newer_model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "%matplotlib inline "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 831,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('cpickles/clean_beer.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 832,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop(columns=['Unnamed: 0'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 833,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 24972 entries, 0 to 24971\n",
      "Data columns (total 133 columns):\n",
      " #   Column           Dtype  \n",
      "---  ------           -----  \n",
      " 0   title            object \n",
      " 1   bstyle           object \n",
      " 2   og               float64\n",
      " 3   fg               float64\n",
      " 4   abv              float64\n",
      " 5   ibu              float64\n",
      " 6   srm              float64\n",
      " 7   Mash pH          float64\n",
      " 8   hops             object \n",
      " 9   others           object \n",
      " 10  water            object \n",
      " 11  notes            object \n",
      " 12  IBU (rager)      float64\n",
      " 13  SRM (daniels)    float64\n",
      " 14  SRM (ebcmorey)   float64\n",
      " 15  SRM (mosher)     float64\n",
      " 16  ABV (alternate)  object \n",
      " 17  yeast            object \n",
      " 18  co2              object \n",
      " 19  chal             int64  \n",
      " 20  citra            int64  \n",
      " 21  warrior          int64  \n",
      " 22  casc             int64  \n",
      " 23  saaz             int64  \n",
      " 24  azacca           int64  \n",
      " 25  colum            int64  \n",
      " 26  mag              int64  \n",
      " 27  perle            int64  \n",
      " 28  simcoe           int64  \n",
      " 29  amar             int64  \n",
      " 30  cashmere         int64  \n",
      " 31  comet            int64  \n",
      " 32  galaxy           int64  \n",
      " 33  apollo           int64  \n",
      " 34  willam           int64  \n",
      " 35  goldings         int64  \n",
      " 36  heller           int64  \n",
      " 37  fug              int64  \n",
      " 38  nugg             int64  \n",
      " 39  hopshot          int64  \n",
      " 40  tett             int64  \n",
      " 41  mosaic           int64  \n",
      " 42  brewerg          int64  \n",
      " 43  northb           int64  \n",
      " 44  cent             int64  \n",
      " 45  chinook          int64  \n",
      " 46  crystal          int64  \n",
      " 47  magnum           int64  \n",
      " 48  ctz              int64  \n",
      " 49  styr             int64  \n",
      " 50  idaho            int64  \n",
      " 51  nelson           int64  \n",
      " 52  exper            int64  \n",
      " 53  hall             int64  \n",
      " 54  noble            int64  \n",
      " 55  dryh             int64  \n",
      " 56  lact             int64  \n",
      " 57  hopcmp           float64\n",
      " 58  clone            int64  \n",
      " 59  juicy            int64  \n",
      " 60  dank             int64  \n",
      " 61  wedding          int64  \n",
      " 62  whisk            int64  \n",
      " 63  trop             int64  \n",
      " 64  award            int64  \n",
      " 65  best             int64  \n",
      " 66  smoke            int64  \n",
      " 67  roast            int64  \n",
      " 68  thick            int64  \n",
      " 69  pine             int64  \n",
      " 70  sticky           int64  \n",
      " 71  foam             int64  \n",
      " 72  hoppy            int64  \n",
      " 73  choco            int64  \n",
      " 74  coffee           int64  \n",
      " 75  straw            int64  \n",
      " 76  bisc             int64  \n",
      " 77  dark             int64  \n",
      " 78  heavy            int64  \n",
      " 79  stfruit          int64  \n",
      " 80  hazy             int64  \n",
      " 81  pheno            int64  \n",
      " 82  ester            int64  \n",
      " 83  banana           int64  \n",
      " 84  bitter           int64  \n",
      " 85  clove            int64  \n",
      " 86  sour             int64  \n",
      " 87  wylon            int64  \n",
      " 88  kveik            int64  \n",
      " 89  omega            int64  \n",
      " 90  us5              int64  \n",
      " 91  us4              int64  \n",
      " 92  chico            int64  \n",
      " 93  yeast_dark       int64  \n",
      " 94  wy1056           int64  \n",
      " 95  wy1272           int64  \n",
      " 96  wlp080           int64  \n",
      " 97  wlp80            int64  \n",
      " 98  vermont          int64  \n",
      " 99  ylager           int64  \n",
      " 100 ykolsch          int64  \n",
      " 101 ipa              int64  \n",
      " 102 saison           int64  \n",
      " 103 lager            int64  \n",
      " 104 stout            int64  \n",
      " 105 kolsch           int64  \n",
      " 106 wit              int64  \n",
      " 107 puree            int64  \n",
      " 108 lovehop          int64  \n",
      " 109 strong           int64  \n",
      " 110 dry              int64  \n",
      " 111 sweet            int64  \n",
      " 112 bugu             float64\n",
      " 113 simsty           object \n",
      " 114 threesty         object \n",
      " 115 foursty          object \n",
      " 116 malt             int64  \n",
      " 117 agg              int64  \n",
      " 118 balance          int64  \n",
      " 119 complex          int64  \n",
      " 120 crisp            int64  \n",
      " 121 smooth           int64  \n",
      " 122 diac             int64  \n",
      " 123 apple            int64  \n",
      " 124 floral           int64  \n",
      " 125 fruity           int64  \n",
      " 126 robust           int64  \n",
      " 127 wood             int64  \n",
      " 128 vanilla          int64  \n",
      " 129 cherry           int64  \n",
      " 130 funk             int64  \n",
      " 131 grapef           int64  \n",
      " 132 solvent          int64  \n",
      "dtypes: float64(12), int64(109), object(12)\n",
      "memory usage: 25.3+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info(verbose = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix, classification_report\n",
    "from sklearn.datasets import make_classification\n",
    "from sklearn.naive_bayes import GaussianNB, BernoulliNB, MultinomialNB\n",
    "from sklearn.svm import LinearSVC, SVC\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import precision_score, recall_score, precision_recall_curve,f1_score, fbeta_score\n",
    "\n",
    "\n",
    "from mlxtend.plotting import plot_decision_regions\n",
    "\n",
    "from imblearn.over_sampling import RandomOverSampler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 409,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "ename": "KeyError",
     "evalue": "\"['gy054'] not in index\"",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-409-4af3135ce1b9>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0my_df\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdf\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'ipa'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'wylon'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'kveik'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'omega'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'us5'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'us4'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'chico'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'yeast_dark'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'wy1056'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'wy1272'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'wlp080'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'wlp80'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'gy054'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'vermont'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'ylager'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m'ykolsch'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdropna\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/pandas/core/frame.py\u001b[0m in \u001b[0;36m__getitem__\u001b[0;34m(self, key)\u001b[0m\n\u001b[1;32m   2906\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mis_iterator\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2907\u001b[0m                 \u001b[0mkey\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2908\u001b[0;31m             \u001b[0mindexer\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mloc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_get_listlike_indexer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mraise_missing\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2909\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2910\u001b[0m         \u001b[0;31m# take() does not accept boolean indexers\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/pandas/core/indexing.py\u001b[0m in \u001b[0;36m_get_listlike_indexer\u001b[0;34m(self, key, axis, raise_missing)\u001b[0m\n\u001b[1;32m   1252\u001b[0m             \u001b[0mkeyarr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindexer\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnew_indexer\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0max\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_reindex_non_unique\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkeyarr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1253\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1254\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_validate_read_indexer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkeyarr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindexer\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mraise_missing\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mraise_missing\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1255\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mkeyarr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindexer\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1256\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/pandas/core/indexing.py\u001b[0m in \u001b[0;36m_validate_read_indexer\u001b[0;34m(self, key, indexer, axis, raise_missing)\u001b[0m\n\u001b[1;32m   1302\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mraise_missing\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1303\u001b[0m                 \u001b[0mnot_found\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0max\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1304\u001b[0;31m                 \u001b[0;32mraise\u001b[0m \u001b[0mKeyError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf\"{not_found} not in index\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1305\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1306\u001b[0m             \u001b[0;31m# we skip the warning on Categorical\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyError\u001b[0m: \"['gy054'] not in index\""
     ]
    }
   ],
   "source": [
    "y_df = df[['ipa', 'wylon','kveik','omega','us5','us4','chico','yeast_dark','wy1056','wy1272','wlp080','wlp80','gy054','vermont','ylager','ykolsch']].dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = y_df.drop(columns =['ipa'])\n",
    "y= y_df['ipa']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, x_test, y_train, y_test = train_test_split(X, y, test_size=0.2)\n",
    "X_train, x_val, y_train, y_val = train_test_split(X_train, y_train, test_size=0.25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 280,
   "metadata": {},
   "outputs": [],
   "source": [
    "ros = RandomOverSampler(random_state=0)\n",
    "X_resampled, y_resampled = ros.fit_sample(X_train,y_train)\n",
    "X_pre = preprocessing.scale(X)\n",
    "X_train_pre = preprocessing.scale(X_resampled)\n",
    "#x_test_pre = preprocessing.scale(x_test_resampled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 300,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_pre = preprocessing.scale(X)\n",
    "X_train_pre = preprocessing.scale(X_train)\n",
    "x_test_pre = preprocessing.scale(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(n_neighbors=23)"
      ]
     },
     "execution_count": 283,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn = KNeighborsClassifier(n_neighbors = 23)\n",
    "knn.fit(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 284,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7676674648492663"
      ]
     },
     "execution_count": 284,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.score(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 285,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7721495671394364"
      ]
     },
     "execution_count": 285,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.score(x_test_pre,y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 286,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, ..., 0, 0, 0])"
      ]
     },
     "execution_count": 286,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.predict(x_test_pre)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 287,
   "metadata": {},
   "outputs": [],
   "source": [
    "k_range = list(range(1, 61))\n",
    "param_grid = dict(n_neighbors=k_range)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "metadata": {},
   "outputs": [],
   "source": [
    "# instantiate the grid\n",
    "grid = GridSearchCV(knn, param_grid, cv=5, scoring='accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 293,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, estimator=KNeighborsClassifier(n_neighbors=23),\n",
       "             param_grid={'n_neighbors': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,\n",
       "                                         13, 14, 15, 16, 17, 18, 19, 20, 21, 22,\n",
       "                                         23, 24, 25, 26, 27, 28, 29, 30, ...]},\n",
       "             scoring='accuracy')"
      ]
     },
     "execution_count": 293,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fit the grid with data\n",
    "grid.fit(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 289,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'GridSearchCV' object has no attribute 'cv_results_'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-289-562ae85c6dc8>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;31m# view the complete results (list of named tuples)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mgrid\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv_results_\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m: 'GridSearchCV' object has no attribute 'cv_results_'"
     ]
    }
   ],
   "source": [
    "# view the complete results (list of named tuples)\n",
    "grid.cv_results_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 290,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'GridSearchCV' object has no attribute 'cv_results_'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-290-6b133c571a35>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      8\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"mean_test_score\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel_grid\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv_results_\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'mean_test_score'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 10\u001b[0;31m \u001b[0mexamine_first_tuple\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgrid\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-290-6b133c571a35>\u001b[0m in \u001b[0;36mexamine_first_tuple\u001b[0;34m(model_grid)\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;31m# examine the first tuple\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mexamine_first_tuple\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel_grid\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 6\u001b[0;31m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"std_test_score\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel_grid\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv_results_\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'std_test_score'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      7\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"rank_test_score\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel_grid\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv_results_\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'rank_test_score'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      8\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"mean_test_score\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel_grid\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv_results_\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'mean_test_score'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'GridSearchCV' object has no attribute 'cv_results_'"
     ]
    }
   ],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\", category=DeprecationWarning)\n",
    "\n",
    "# examine the first tuple\n",
    "def examine_first_tuple(model_grid):\n",
    "    print(\"std_test_score\", model_grid.cv_results_['std_test_score'])\n",
    "    print(\"rank_test_score\", model_grid.cv_results_['rank_test_score'])\n",
    "    print(\"mean_test_score\", model_grid.cv_results_['mean_test_score'])\n",
    "    \n",
    "examine_first_tuple(grid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 291,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'GridSearchCV' object has no attribute 'best_score_'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-291-3ba4bb78f926>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mgrid\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbest_score_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgrid\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbest_params_\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m: 'GridSearchCV' object has no attribute 'best_score_'"
     ]
    }
   ],
   "source": [
    "grid.best_score_, grid.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# examine the best model\n",
    "def get_best_model(model_grid):\n",
    "    best_k = model_grid.cv_results_['rank_test_score'][0]\n",
    "    print (best_k, model_grid.cv_results_['mean_test_score'][best_k], model_grid.cv_results_['std_test_score'][best_ka])\n",
    "\n",
    "get_best_model(grid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(n_neighbors=15)"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn = KNeighborsClassifier(n_neighbors = 15)\n",
    "knn.fit(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7671762755571928"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.score(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.767544667526248"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.score(x_test_pre,y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, ..., 0, 0, 0])"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.predict(x_test_pre)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve, roc_auc_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_test, knn.predict_proba(x_test_pre)[:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KNN ROC AUC score =  0.5072376777545388\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "plt.plot(fpr, tpr,lw=2)\n",
    "\n",
    "plt.plot([0,1],[0,1],c='violet',ls='--')\n",
    "plt.xlim([-0.05,1.05])\n",
    "plt.ylim([-0.05,1.05])\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC Score', fontweight=\"bold\", fontsize=20);\n",
    "plt.legend(['KNN'])\n",
    "print(\"KNN ROC AUC score = \", roc_auc_score(y_test, knn.predict_proba(x_test_pre)[:,1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Naive Bayes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7690182354024683"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bnb = BernoulliNB()\n",
    "bnb.fit(X_train_pre, y_train)\n",
    "bnb.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "fprbnb, tprbnb, thresholdsbnb = roc_curve(y_test, bnb.predict_proba(x_test_pre)[:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BNB ROC AUC score =  0.5308136836375097\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEbCAYAAADXk4MCAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAA/pElEQVR4nO3dd5xU9fX/8deZ2ZntLF3qskuXDiIqioLYsBeM7RujyVejv6iJSbAm1lgxsUXj18SWokasqKhREVFBA+guVZYqLCAdlm1Tz++Pe8FhmV0G3dk25/l4zIO5cz/3zueyu3Pm3vu57yuqijHGmNTlaewOGGOMaVxWCIwxJsVZITDGmBRnhcAYY1KcFQJjjElxVgiMMSbFWSEwxpgUZ4XANAgRKRARjfOIikiFiCwTkX+KyFH7WU+WiFwpItNEZJ2IVItIuYisEJEXROQMEZH9rMMjIqeJyD9EpEREykSkyl3HhyLyaxHpfoDbd7aIvCUiG0QkKCK7RGSNiHwuIn8VkZ8fyPqMaUhiF5SZhiAiBcCqBJoq8DNVfSbOOsYALwBd97OOWcB5qloaZx19geeBQ/azjo9VdWwC/UVE/gb8bD/Ndqpq60TWZ0xDs0JgGkScQjAX+DeQARwBnBwzbzPQSVWjMcsfCnwMZMa0+w/wqfvaacCgmHklwOGquj1mHb2Bz4COMe2+Ad4E1gN5OAViHPBpIoVARE4A3ot56St3eifQFhgCHAMEmmIhEJFWqlrW2P0wjUxV7WGPpD+AApxv+7sfz9aY/3mN+QfFzBOguMb8/62xvAd4rkabR2u0+bjG/CcBX5y+dgV+nuB2/SlmfcuBtDhtcoDTalm+FTAJ+ATYCgSBjTh7NbfU8v/4ILAQKAeqcYrZi8CYOO0vqbHNvYHrgCVAAJhRo/0JwBRgrTu/DPgv8Fsgq7F/j+yRnEejd8AeqfFIoBC8EjMvAvhj5o2psezsWt6jvfvhuLtdBZDuzju0xjqKAG89bNfDMevcAvQ5gGWHAGtq9Cv2saNG+9OAXXW0V+DuGsvULAQza0zPcNuJWxjrWvd8oGNj/y7Zo/4faRjTiEQkAxgNHB/z8suqGoyZPqbGYi/GW5eqbhGRD4Az3JeygJE4h4PG12j+rKpGvnfHv/NlzPN2wFIRWYBz6Osr4BNVLa65kIhkA9PY+3zHHGC6+3yE2/fd7QtxDqXtPjRWCTyLcwjqPKCn+/qNIrJQVZ+vpb9jcPYGpgLRmPX9Brgspt07wGygA3AxzmGzwcA/cfYaTEvS2JXIHqnxYN89gtoerwF5NZZ9vEabM+p4n4dqtD3Xff2xGq+fVE/blca+h7VqPr4Gzqyx3C9qtHkM95xdTJteMc//WKP9cTHz2gLbYuYVx8y7pMZys4GMGu/jATbFtHm8xvwJNdYxrLF/n+xRvw8bPmqakvnA71V1Z2N3JFGqGsbZ27gd2GeUkqsf8KqInBbz2tE12tys7qduzLpXxEzGDqstVdUPYtptA96ImT9ERHJr6csDqlodp38dYqavjB3ii7PnEqvOIb6m+bFCYBrLXJyTlk/inPAE55j5JyLSq0bb9TWmC+pYb815u5et+SHdP6FeJkBVK1T1NlXtjvOhejHwfzjfsncT4Ncx021jnm9X1R37eZvY9t/GmV/ztTa1rOfr/aw7ER3238Q0J3aOwDSWRao6GUBE3sE5JATQGniUvYeTflxj2fNwTtLuRUTasfe5gCqcggPwAXB3zLyfiMgjGjNEtT6oagnO0NV/iMgknA/eLu7sHjFNt8U8byMirfdTDGLbd4ozv+Zr22tZT8V+1g3wMvBFHX2ZXcc80wzZHoFpdKr6OnuPxZ/gXjy226c4h412O0JELo1dh4h4gAdwhmru9pSqBtz3mIMzRHO3YcCfRWSfL0Mi0jXRK4FF5Cci8v9EpHWc2QGc4aC7bY15PrNG2ztqXhHtniDe7bOY591E5LiYdm357gQ5wHxV3ZVI/11Lca7d2K0t8JCqPhD7wDmPsVlVP4u7FtNs2QVlpkHEuaDsOVW9JGb+aPb+sPtQVWM/7EYBM9j7grJ3ccbbp+MMrRwSM28ZzgVl22LW0cd9j9hDG6txRtBsYO8Lyj7TxC4ouw24FQi56/4K50M1FzilRp/uUNVb3eVycPYWYkcNfYEzakhxRuiMVtX2bvuewCKcC/DAGTX0DM44/9hRQwA/VtV/ustd4rbbrVBVV8fZjuuA+2Je+hp4C6d47b4w7mggU1XrjPAwzVBjn622R2o82M91BG6bj2q0OarG/KOBdTXaxHvMArrX0o/+OEM+97eOGQlu120JrEtxhobm1lh2KAd2HcGZOId26nqf+2ssc0mN+QW1bIcAf01kWxr7d8ke9f+wcwSmKbkTGBszfRuwZ69AVWe6WUE/wdkDGIbzbTWMc2L2C5xrDKZqLcf+VfVrERkJnA6cC4zCOb7uwzmxvBp4G+fq2kQ8hHPV81ici9Y64URY+HGO0y/CuVjur7r3tRGoarGIDAZ+7vZnAM6exE5gJTVG66jq6277X+Jcd9ED5zzfJpzi9xdVnZFgv/eiTjW4TEReBP4XOBzojLOnswFnb+4/OHEcpoWxQ0PGGJPi7GSxMcakOCsExhiT4qwQGGNMirNCYIwxKa7ZjRpq3769FhQUNHY3jDGmWZk3b94WVY0bD9LsCkFBQQFz587df0NjjDF7iMg3tc2zQ0PGGJPirBAYY0yKs0JgjDEprtmdI4gnFApRWlpKdXXN+220DBkZGXTr1g2fz9fYXTHGtEAtohCUlpaSm5tLQUEBNZJ8mz1VZevWrZSWllJYWLj/BYwx5gAl7dCQiDwtIptEZGEt80VEHhGR5SIyX0RGfN/3qq6upl27di2uCACICO3atWuxezvGmMaXzHMEzwIn1TF/AtDHfVwO/OWHvFlLLAK7teRtM8Y0vqQdGnIjgwvqaHIG8Hc3/vZzEWktIp1VdUOy+mSMMc3NN1sr+GjxJhYs3kFOJx+3nzGo3t+jMUcNdQXWxkyXsvfdmvYQkctFZK6IzN28eXO8Jo3O6/UybNgwhg4dyogRI5g1axYAq1evRkR49NFH97S96qqrePbZZwG45JJLKCwsZNiwYfTv35/bb7+9MbpvjGlCguEoL88r5aSHZnL5H+fQ78M0Lv+2Mx8v2kwybh3QmIUg3vGOuFuoqk+q6khVHdmhQ9wrpBtdZmYmRUVFFBcXc88993DjjTfumdexY0cefvhhgsFg3GUnT55MUVERRUVFPPfcc6xatSpuO2NMy1YRCPO3T1ZyzOSPuGnKfI7f2oZ/Zgyhqz+DjSOVl68ZnZRDxY05aqgU6B4z3Q3nDlHNXllZGW3atNkz3aFDB4488kiee+45LrvsslqX231CODs7O+l9NMY0DapK0dodfLBkI//8fA07q0IAPNNqMMMiuXiH+Gh3Qhb5mcn73t6YhWAqcJV7a7zDgJ31cX6g4Ia3f3DH4ll97yl1zq+qqmLYsGFUV1ezYcMGpk+fvtf8G264gQkTJvDTn/50n2UnTZrEH/7wB5YvX84111xDx44d67XvxpimadaKLTzw3lK+XLMDgCw8HJ7flv8d25MjM9sgKvh6Jv/6oaQVAhF5Aec+ru1FpBS4Fee+sKjqEzj3Yz0ZWA5UApcmqy8NYfehIYDZs2dz8cUXs3DhdyNnCwsLGTVqFM8///w+y06ePJmJEydSXl7O+PHjmTVrFqNHj26orhtjGljR2h088N5SPl2+BYDWWT5+2rU75+88iNz+mWQOyGzQ/iRz1NAF+5mvwC/q+3339829IRxxxBFs2bKFmie2b7rpJiZOnMjRRx8dd7mcnBzGjh3Lp59+aoXAmBZoyYYy/vifEj5YshGA3Iw0rjq8F+dVHYQuDONp58HXp+ETBFrElcVNzddff00kEqFdu3ZUVlbueb1///4MGDCAt956i1GjRu2zXDgc5osvvuDqq69uyO4aY5Js1ZYKHny/hDfnr0cVMn1eLj2ygMt69EDeCaBVYTKOyiBjTAaS1vDXDVkhqCe7zxGAc/Lnueeew+v17tPu5ptvZvjw4Xu9tvscQTAYZPz48Zx99tkN0WVjTJKt21HFIx8s4+UvS4lEFb/Xw4WH5fOLcb3pkJtOZGOEitYhsi7MIq1T430cWyGoJ5FIJO7rBQUFe50rGDp0KNFodM/07usJjDEtx+ZdAR77aDnPf7GGYCSK1yOcf2h3rj62N+3XeIl8FoGTwHuQl9xLcxs9PcAKgTHG1JMdlUH+b+ZKnv1sNVWhCCJw+tAuXHt8X/K9GVROq6RyZYC0/DQ0pIhPGr0IgBUCY4z5wcoDYZ75dBVPfrKSXdVhAI4fcBC/OaEv/TrmEpgboGx6GQhkTcjCf4i/SRSA3VpMIVDVJvUfW5+ScUm5MeaHqw5F+Ofn3/D4jBVsq3CSA47q3Z7fnNCX4fnORaXR8ijVM6pJ65FG9snZePKa3v3AWkQhyMjIYOvWrS0yinr3/QgyMjIauyvGGFcoEuWluWt59MPlfFvmJAKMyG/Nb0/sx+he7dGIEigO4B/ix5PjIfeyXDytPU3286lFFIJu3bpRWlq6z7j9lmL3HcqMMY0rElWmFq/jwfeXsWabMzR8QOdW/PbEvozr1xERIbwhTOXUSiKbInhyPPh6+fC22XcEYVPSIgqBz+ezu3cZY5JGVXlv0bf88T8lLNtUDkDPDtn85vh+TBjUCY9H0JBSObOSwOwAki1kn5uNr1fzuL1siygExhiTDKrKzGVOHtCCdTsB6No6k18d14ezhnclzfvd8f7yl8oJrwzjH+4n87hMPBlN71xAbawQGGNMDSUbd/HIh8tYtL6MVVsqAOiQm87Vx/bmvEO7k57mHOrRgIIXJE3IOCoDRoOvsHnsBcSyQmCMMa6dVSEefL+Ef3z+DZGoM1qvdZaPK4/pxcVHFJDp/+5Yf2hZiIppFaQPTifz2Ex8PZpfAdjNCoExJuVFosqUuWu5/72lbKsI4hH48eE9OH9Ud3p1yCHD910BiFZGqfpPFcEFQTztPfj6Nt8CsJsVAmNMSpv3zTZum7p4zzmAUYVtue20gQzo0mqftqGVISpeq0CrlYwxGWQc1TghcfXNCoExJiVtLKvmvne+5tWv1gHQOS+Dm04+mFOHdK51vL8nx4OnnYfsCdl4D2raQ0IPhBUCY0xKCYQjPPPZah79cBkVwQj+NA8/P7onV47tRZZ/749EVSVYFCTybYSsCVl4O3rJ/Unjh8TVNysExpiU8dHXm7jjrcV7RgIdP+Agfn/KAPLbZe3TNrI9QuVblYRXh0nr0bRC4uqbFQJjTIu3aksFd761mOlfbwKci8FuPW0gx/TtsE9bjSqB/wao+qgKPJB1Shb+4U0rJK6+WSEwxrRYFYEwj05fzlOfriQUUXLS0/jVcX24+IgC/GnxL/jSSqV6ZjW+Qh9ZJ2fhadV8Lgz7vqwQGGNaHFXljaL13PPOEjaWBQA495BuTDqpHx1z9w1w1IgSXBDEP9QNibs8F09e0w2Jq29WCIwxLcrCdTu5deoi5n2zHYCh3Vtz22kD9sRC1xReH6ZiagXRzVE8uW5IXOuWMyIoEVYIjDEtwtbyAA/8p4QX56xBFdrn+Ln+pP6cM6IbHs++3+w1pFTNqCLwRQDJEbLPaz4hcfXNCoExptkKRaLMXb2dD5ZsZMrctZRVh0nzCJccWcA1x/WhVUbtH+zl/y4nvCqMf4SfrPFZSEZqHAaKxwqBMabZiUaVF+as4f53l7KzKrTn9TF92nPraQPo3TE37nJarZDmhsSNyYCjwFeQmnsBsawQGGOalTVbK7n+lfnMXrkVgF4dshnXryPj+ndkdK/a71IYLAlSOa3SCYkb37xD4uqbFQJjTLMQjSrPzV7N/e8upSoUoW22n9tPH1hnJARAtCJK5XuVhBaF8Hb04jvYCkBNVgiMMU3eys3lXPfyfOa6I4FOG9qF204bQLuc9DqXC60IUfG6GxJ3TAYZR2Yg3tQ9F1AbKwTGmCYrHIny1Ker+NP7JQTCUTrkpvOHMwdx4sBOCS3vyfXgbe/dkxNk4rNCYIxpkko27mLSlGKKS5146HNGdOP3px5M6yx/rcuoKsGvgoS/DZN9cvaekDhTNysExpgmJRSJ8sSMFTwyfRmhiNI5L4O7zxrMuP4d61wuss0NifsmTFrBdyFxZv+sEBhjmoxF63cyacp8Fm8oA+CCUd258eSD67weQKNK4IsAVTOqwAtZp2bhH9ayQ+LqW1ILgYicBDwMeIG/qeq9NebnAf8E8t2+PKCqzySzT8aYpicQjvDY9OU8PmMF4ajSrU0m9549hKP6tN/vslqpVH9aja+nj6wJqRESV9+SVghExAs8BhwPlAJzRGSqqi6OafYLYLGqniYiHYClIvIvVQ0mq1/GmKaleO0OJr1cTMnGcgB+ckQPrjupP9nptX88aVgJzg/iH+6GxF2WWiFx9S2ZewSjgOWquhJARF4EzgBiC4ECueL89HKAbUA4iX0yxjQR1aEID35Qwl9nriSqUNAui/vOGcJhPdvVuVx4XZiKN92QuLzUDImrb8ksBF2BtTHTpcBhNdr8GZgKrAdygfNUNVpzRSJyOXA5QH5+flI6a4xpOPO+2cakl+ezcnMFHoHLxhTy6+P7kemv/QNdgzEhca2EnPNzUjYkrr4lsxDE20fTGtMnAkXAsUAv4H0R+URVy/ZaSPVJ4EmAkSNH1lyHMaaZqAyGmfzeUp6dtRpV6N0xh/snDmFELRHRscpfckLi0g9xIiIk3Q4D1ZdkFoJSoHvMdDecb/6xLgXuVVUFlovIKqA/8N8k9ssY0whmrdjCDa8sYM22Srwe4YqxPbn62D5k+GrfC4hWRxGvID4h4+gMGINlBCVBMgvBHKCPiBQC64DzgQtrtFkDjAc+EZGDgH7AyiT2yRjTwMoDYe6ZtoR/fbEGgP6dcpk8cSiDu+XVuVxwaZDKdyrxD3Zion35VgCSJWmFQFXDInIV8B7O8NGnVXWRiFzhzn8CuBN4VkQW4BxKul5VtySrT8aYhvVxyWZuenUB63ZUkeYRrjq2N/9vbO9a7xcMbkjcu5WEFjshcf6Da7+S2NSPpF5HoKrTgGk1Xnsi5vl64IRk9sEY0/B2VoW46+3FvDS3FIDBXfO4f+IQDu7cqs7lQsvdkLigkjE2g4zRFhLXEOzKYmNMvfpwyUZuem0BG8sC+L0efnV8Hy4f05M07/4v9PK08uDt6IbEdbAhoQ3FCoExpl5srwhy+5uLeL3IGRMyPL81kycOqfVuYeCGxM0LEt4YJvsUNyTuYguJa2hWCIwxP9g7Czbw+zcWsqU8SIbPw29P6MelRxbijXPT+N0iW92QuDVh0nqmoWFF0uwwUGOwQmCM+d427wpw69SFTFvwLQCjCtty/zlDKGifXesyGlUCswNUfVyF+ISs07PwD7GQuMZkhcAYc8BUlanF67lt6iK2V4bI8nu5cUJ/LjqsB5469gLADYmbVY2vtxsSl2shcY3NCoEx5oBsLKvm5tcW8MGSTQAc1bs995w9mO5ts2pdRsNKsDiIf4QTEtfq8lZ48qwANBVWCIwxCVFVpswr5c63FrOrOkxueho3n3Iw5x3avc7DOuFSNyRuSxRPGw++nj4rAk2MFQJjzH6t21HFja8uYGbJZgDG9evA3WcPpnNeZq3LaFCp+qiKwH8DePI85FyYg6+nXR3cFFkhMMbUKhpVXpizhnumfU15IExepo9bTxvAWcO77vfk7p6QuEPTyRxnIXFNmRUCY0xca7ZWcv0r85m9cisAJw48iDvPHETH3Ixal4lWRZE0JyQu8+hMOBrS8u1jpqmzn5AxZi/RqPLc7NXc/+5SqkIR2mb7ueOMgZwyuHOdewHBJUEq33VD4o7LsgLQjCT0kxKRTCBfVZcmuT/GmEa0cnM51708n7nfbAfgtKFduO20AbTLSa91mWh5lMp3Kgl9HcJ7kBf/QAuJa272WwhE5DTgAcAPFIrIMOAOVT09yX0zxjSQcCTKU5+u4k/vlxAIR+mQm84fzhzEiQM71blcaHmIitcq0JCSMS6DjCMsJK45SmSP4Dac+w/PAFDVIhEpSF6XjDENqWTjLiZNKaa4dCcA54zoxu9PPZjWWfv/Zu/J8+Dt5IbEtbeQuOYqkUIQVtWddvm3MS1LKBLliRkreGT6MkIRpXNeBnefNZhx/TvWuoyqEpgbILIxQvap2Xg7eMn9sYXENXeJFIKFInIh4BWRPsA1wKzkdssYk0yL1u9k0pT5LN7g3B78glH53Hhyf1pl1D7OP7IlQsVbFUTWRkjrZSFxLUkiheBq4GYgADyPc8exO5PZKWNMcgTCER6bvpzHZ6wgHFW6tcnkvnOGcGTv9rUuoxGlenY11TOrLSSuhUqkEJyiqjfjFAMARORcYErSemWMqXfFa3cw6eViSjaWA3DJ6AImndiP7PS6Pwa02kkL9fX1kXVSFp4ci4doaRIpBDey74d+vNeMMU1QdSjCgx+U8NeZK4kqFLTL4v6JQxlV2LbWZTSsBIoCpB+SjifbQ6uft8LTygpAS1VrIRCRCcDJQFcReSRmVisgnOyOGWN+uLmrt3HdK/NZubkCj8BlYwr59fH9yPTXPsInvMYNidsWxdvW64TEWRFo0eraI1gPzAVOB+bFvL4LuDaZnTLG/DCVwTCT31vKs7NWowq9O+Zw/8QhjMhvU+syGlCqplcRmBvA09pDzkUWEpcqai0EqloMFIvI86oaasA+GWN+gFkrtnDDKwtYs60Sr0e4YmxPrj62Dxm+usf5l79UTnh1mPRRbkic304Gp4pEzhEUiMg9wABgT9qUqvZMWq+MMQesPBDmnmlL+NcXawDo3ymXB84dyqCuebUus1dI3NhMEEjrZhlBqSaRn/gzwK3Ag8A44FLAvioY04R8XLKZm15dwLodVfi8wlXj+nDl2F7402o/th9c7IbEDXFD4rpbAUhVifzkM1X1QxERVf0GuE1EPsEpDsaYRrSzKsRdby/mpbmlAAzumsfkc4fQv1OrWpeJ7nJD4paG8Hb24h9kIXGpLpFCUC0iHmCZiFwFrANqvwbdGNMgPlyykZteW8DGsgD+NA+/Oq4Pl4/pSZq39r2A0LIQFa9XoGElc3wm6YenI/u52bxp+RIpBL8CsnCiJe7EOTz0kyT2yRhTh+0VQW5/cxGvF60HYHh+ayZPHELvjvvP/PG09uDt4iXrpCy87SwkzjjqLAQi4gV+pKqTgHKc8wPGmEbyzoIN/P6NhWwpD5Lh8/DbE/px6ZGFeGv5Vq9RJTAnQGRThOzT3JC4iywkzuytzkKgqhEROcQ9P6AN1SljzN427wpw69SFTFvwLQCjCtty/zlDKGifXesykc1uSFxphLTeFhJnapfIoaGvgDdEZApQsftFVX01ab0yxgBO7PPU4vXcNnUR2ytDZPm93DihPxcd1gNPbXsBEaV6VjXVn1QjfiHrzCz8gywkztQukULQFtgKHBvzmgL7LQQichLwMOAF/qaq98ZpMxZ4CPABW1T1mAT6ZEyLt7GsmptfW8AHSzYBMKZPe+4+azDd22bVuZxWK4EvAvj6uSFx2RYPYeq230Kgqt/rvIB7fuEx4HigFJgjIlNVdXFMm9bA48BJqrpGRGw0kkl5qsqUeaXc+dZidlWHyU1P43enHsyPRnav9Vu9htyQuJExIXG5VgBMYpJ5BckoYLmqrgQQkReBM4DFMW0uBF5V1TUAqropif0xpslbt6OKG19dwMySzQCM69eBu88eTOe8zFqXCX0TovKtSickrr0XX6HPioA5IMksBF2BtTHTpcBhNdr0BXwiMgPIBR5W1b/XXJGIXA5cDpCfn5+UzhrTmKJR5YU5a7hn2teUB8LkZfq49bQBnDW8a+17AQGl6sMqAvPckLj/ycFXaCFx5sAlsxDE++2tOfIoDTgEGA9kArNF5HNVLdlrIdUngScBRo4caaOXTIuyZmsl178yn9krtwJw4sCDuPPMQXTMzahzuT0hcYelkznWQuLM97ffQiAiBwF3A11UdYKIDACOUNWn9rNoKdA9ZrobTrR1zTZbVLUCqBCRmcBQoARjWrhoVHlu9mruf3cpVaEIbbP93HHGQE4Z3LnWvYBoZRTxuSFx45zDRRYSZ36oRA4kPotzn+Iu7nQJztXG+zMH6CMihSLiB84HptZo8wYwRkTSRCQL59DRkgTWbUyztmJzOT/6v9nc/uZiqkIRTh/ahfevPZpTh3SJWwRUleDCIGV/KaPq4yrAKQBWBEx9SOS3qL2qviQiNwKoalhEIvtbyG13FU4R8QJPq+oiEbnCnf+Eqi4RkXeB+UAUZ4jpwu+9NcY0ceFIlKc+XcWf3i8hEI7SITedu84cxAkDO9W6TLTMDYkrCeHt4iV9cHoD9tikgkQKQYWItMM9vi8ihwM7E1m5qk4DptV47Yka05OByQn11phmrGTjLiZNKaa41PnzOWdEN245dQB5WbWf4A2WBKl4vQIikHlcJumHWUicqX+JFILf4BzS6SUinwEdgIlJ7ZUxLUgoEuWJGSt4ZPoyQhGlc14Gd589mHH99n/ZjLetl7RuaU5IXFsLiTPJkcgFZfNE5BigH85IoKV260pjErNo/U4mTZnP4g1lAFwwKp8bT+5Pq4z4ewEaVQL/DRDZGCH7jGy87b3kXmghcSa5Ehk1VAz8G/i3qq5IfpeMaf4C4QiPTV/O4zNWEI4q3dpkct85Qziyd/tal4lsckPi1kXw9fFZSJxpMIkcGjodOA94SUSiOEXhpd1XAxtj9la8dgeTXi6mZGM5AJeMLmDSif3ITo//56YRpfozNyQuQ8g+KxvfQJ+FxJkGk8ihoW+A+4H7RaQP8HvgPpyRQMYYV3UowoMflPDXmSuJKhS0y+L+iUMZVdi2zuW02jkc5B/gJ/OETAuJMw0uoUHIIlIA/AhnzyACXJfEPhnT7MxdvY3rXp7Pyi0VeAQuP7on1x7Xl0x//O9LGlICXwZIP9RC4kzjS+QcwRc4EdFTgHN3h8gZY6AyGGbye0t5dtZqVKF3xxwmTxzC8Pw2tS4TWu2GxG2P4u1oIXGm8SWyR/ATVf066T0xppmZtWILN7yygDXbKvF6hCvG9uSa8X1IT6tlL6BaqfywkuCXQTxtPOT8OAdfgYXEmcZXayEQkf9R1X8CJ4vIyTXnq+qfktozY5qo8kCYe6Yt4V9fOOMl+nfK5YFzhzKoa17dy71UTnhNmPQj0sk8JhPx2clg0zTUtUew+2ao8QYxWwKoSUkfl2zmplcXsG5HFT6vcNW4Plw5thf+tPiHdqIVUcTvhsQdmwkeSOti+UCmaan1N1JV/899+oGqfhY7T0SOTGqvjGlidlaFuOvtxbw0txSAwV3zmHzuEPp3ahW3vaoSWhii8r1K/MP8ZB2XZQFxpslK5DfzUWBEAq8Z0yJ9uGQjN722gI1lAfxpHq49ri+XjSkkzVvLXkBZlMpplYSWhfB29ZI+xELiTNNW1zmCI4DRQAcR+XXMrFbYNQQmBWyvCHL7m4t4vci5jcaI/NbcP3EovTvm1LpMcKkbEqeQeUIm6YdaSJxp+uraI/ADOW6b2PMEZVjonGnhpi3YwC1vLGRLeZAMn4dJJ/bnktEFePfzoe5t5yUt3w2Ja2Pfl0zzUNc5go+Bj0XkWffqYmNavM27Atw6dSHTFnwLwGGFbbnvnCEUtM+O216jSuALNyTuTDck7gILiTPNS12Hhh5S1V8BfxaRfUYJqerpyeyYMQ1JVZlavJ7bpi5ie2WILL+XGyf056LDeuCpZS8gvDFM5ZuVRDZE8PW1kDjTfNV1aOgf7r8PNERHjGksG8uqufm1BXywZBMAY/q05+6zBtO9bVbc9hpWqj+tpvqzaiRTyD4nG9/BFhJnmq+6Dg3Nc//9ePdrItIG6K6q8xugb8YklaoyZV4pd761mF3VYXLT0/jdqQfzo5Hd6/xQ14ASmBfAP9ANicuyeAjTvCWSNTQDJ4o6DSgCNovIx6r667qWM6YpW7ejihtfXcDMks0AHNu/I3edNYjOeZlx22vQDYkbFRMSl2MFwLQMiVxHkKeqZSLyv8AzqnqriNgegWmWolHlhTlruGfa15QHwuRl+rjt9AGcOaxrrXsBoVVuSNyOKN6D3JA4KwKmBUmkEKSJSGecGOqbk9wfY5JmzdZKrn9lPrNXbgXgpIGduOPMgXTMzYjbPlodper9KoJFQTxtPeRcnIOvh4XEmZYnkUJwB/Ae8JmqzhGRnsCy5HbLmPoTjSrPzlrN5PeWUhWK0C7bzx1nDOLkwZ3qPBdQ8VKFExI3Op3Moy0kzrRcidyhbArOvQh2T68Ezklmp4ypLys2l3P9y/OZ+812AE4f2oVbTxtAu5z4sQ/Rcjckzi9kjndD4jpbRpBp2RI5WdwNJ1voSJzU0U+BX6pqaZL7Zsz3Fo5EeerTVfzp/RIC4SgdctO568xBnDCwU9z2qkpwQZCq/1ThH+on6/gs0rpaATCpIZHf9GeA54Fz3en/cV87PlmdMuaHKNm4i0lTiiku3QnAOSO6ccupA8jLin98P7ozSsW0CsLLw3i7eUkfZiFxJrUkUgg6qOozMdPPisivktQfY763UCTKEzNW8Mj0ZYQiSue8DO4+ezDj+nWsdZm9QuJOzCR9pIXEmdSTSCHYIiL/A7zgTl8AbE1el4w5cIvW72TSlPks3lAGwIWH5XPjhP7kZsTfC1BVRARvOy++Hj4yT8rE29pC4kxqSqQQ/BT4M/CgO/2Z+5oxjS4QjvDY9OU8PmMF4ajSrU0m950zhCN7t4/bXqNKYHaAyKYI2Wc5IXE559ceK21MKkhk1NAanCuLjWlSitbu4LqXiynZWA7AJaMLmHRiP7LT4/9ah791Q+K+jeDrZyFxxuyWyKihnsDDwOE4o4ZmA9e6w0iNaXDVoQgPvl/CXz9ZSVShsH02950zhFGFbeO217BS/Uk11bPckLiJ2fgP9jdwr41puhI5NPQ88Bhwljt9Ps75gsOS1SljajN39Taue3k+K7dU4BG4/OieXHtcXzL9tR/f14CTE+Qf5IbEZVo8hDGxEvmLEFX9h6qG3cc/cfYM9r+gyEkislRElovIDXW0O1REIiJidz4zcVUGw9z+5iLO/b/ZrNxSQe+OObxy5WhuOvnguEVAg0r17Go0qk5I3BWtyD4j24qAMXEkskfwkfsh/iJOATgPeFtE2gKo6rZ4C4mIF2dP4nigFJgjIlNVdXGcdvfhxFgYs49ZK7ZwwysLWLOtEq9HuGJsT64Z34f0tPh7AaEVISrfriS6M4q3sxdfgQ9PthUAY2qTSCE4z/335zVe/ylOYehZy3KjgOW7zyWIyIvAGcDiGu2uBl4BDk2kwyZ1lAfC3DNtCf/6Yg0A/Tvl8sC5QxnUNS9u+2iVGxJXHMTTzkPuJbmkdberg43Zn0RGDRV+z3V3BdbGTJdS47yCiHTFOfdwLHUUAhG5HLgcID8//3t2xzQnH5ds5qZXF7BuRxU+r3D1sX244phe+NNq/2Zf8VIF4bVhMo7MIOPoDBsRZEyCkvl1Kd5fYc1zCw8B16tqpM47Qqk+CTwJMHLkyITOT5jmaWdViLveXsxLc50oqyHd8rh/4hD6d2oVt/1eIXHHZYIX0jrZXoAxByKZfzGlQPeY6W7A+hptRgIvukWgPXCyiIRV9fUk9ss0UR8s3sjNry9gY1kAf5qHa4/ry2VjCknz7rsXoKoE57shccMsJM6YHyKZfzlzgD4iUgiswxl2emFsg9jDTiLyLPCWFYHUs70iyG1vLuKNIud7woj81tw/cSi9O8a/4jeyI0Ll25WEV4ZJ655G+nALiTPmh0jkgjIBLgJ6quodIpIPdFLV/9a1nKqGReQqnNFAXuBpVV0kIle485/44d03zd20BRu45Y2FbCkPkuHzMOnE/lwyugBvLcFvwa/dkDiBzJPckLg6DisaY/YvkT2Cx4EozgndO4BdJDjKR1WnAdNqvBa3AKjqJQn0xbQQm3cFuHXqQqYt+BaAwwrbct85Qyhonx23/Z6QuA7OPYMzT7SQOGPqSyKF4DBVHSEiXwGo6nYRsevzzfeiqkwtXs9tUxexvTJEtt/LDScfzEWj8vHE2QvQiHNhWGRThJyzc/C285JznoXEGVOfEikEIfeiLwUQkQ44ewjGHJCNZdXc/NoCPliyCYAxfdpzz9mD6dYmK2778AY3JG5jBN8AC4kzJlkSKQSPAK8BHUXkLmAi8Luk9sq0KKrKlHml3PnWYnZVh8nNSOP3pwzg3JHd4h7f15BSPbOa6tnVSLaQfW42/v62E2pMsiRyQdm/RGQeMB7n2oAzVXVJ0ntmWoR1O6q48dUFzCzZDMCx/Tty91mD6ZSXUesyGlICRQH8Q/1kHmchccYkWyKjhvKBSuDN2Nfc+xQYE1c0qjz/3zXcM20JFcEIeZk+bjt9AGcO6xp/LyCgBOYFSD88HU+Wh1ZXtsKTZQXAmIaQyKGht3HODwiQARQCS4GBSeyXacbWbK3k+lfmM3ulc0fTkwZ24o4zB9IxN/5eQGh5iMppbkhcFzckzoqAMQ0mkUNDg2OnRWQE+wbQGUM0qjw7azWT31tKVShCu2w/d5wxiJMHd4q7FxCtdEPi5gfxtPeQe2kuad3s6mBjGtoB/9Wp6pciYkmhZi8rNpdz/cvzmfvNdgBOH9qFW08bQLuc2q/6rZhSQbg0TMaYDDKOspA4YxpLIucIfh0z6QFGAJuT1iPTrIQjUZ76dBV/er+EQDhKx9x0/nDmIE4Y2Clu++iuKJLuhsQdnwkeC4kzprEl8heYG/M8jHPO4JXkdMc0JyUbdzFpSjHFpTsBmHhIN35/ygDysnz7tFVVgsUxIXEnZJHWxQqAMU1BnX+J7oVkOao6qYH6Y5qBUCTKEzNW8Mj0ZYQiSue8DO45ezBj+3WM2z6y3Q2JWxUmLT+N9EMsJM6YpqTWQiAiaW5w3IiG7JBp2hat38mkKfNZvKEMgAsPy+fGCf3Jzdh3LwAguCRIxRtOSFzWyVn4R/gtJM6YJqauPYL/4pwPKBKRqcAUoGL3TFV9Ncl9M03M3z5Zyb3vfE04qnRrk8l95wzhyN7t47bdExLX0Yuvl4+sE7Lw5NmQUGOaokQO0rYFtuKkj+6+nkABKwQp5KOlm/jD284F5ZeMLmDSif3ITt/310cjSvWsaiKbI2Sfle2ExJ1rIXHGNGV1FYKO7oihhXxXAHaz20WmiEhUefjDZTw6fRkA5x/andtOj38tYXi9GxK3KYJvoA8iJPfWR8aYelHXn6kXyCGxew+bFmhLeYBfv1TMzJLNeAQuPbKQ357Yd592GlKqPq4i8HkAyRGyf5SNv5+FxBnTXNRVCDao6h0N1hPTpKzZWsn/PPUFa7ZV0ibLxyMXDGdMnw5x22rIGRrqH+aGxGXYuQBjmpO6CoEN7UhRS7/dxWV/n8uabZUM7prHXy8euU9aqAaUwNwA6UdYSJwxzV1dhWB8g/XCNAmqyt9nf8Nd05YQDEcZ0LkVz1922D5DQ0PLQlRMq0B3Kd6uFhJnTHNXayFQ1W0N2RHTuLaUB7ju5flM/9q5e9j5h3bnltMGkOX/7lckWhGl6j9VBBcG8XTwkDMxh7SudjbYmObO/ooNH5ds5jcvFbOlPEBepo97zx7MhMGd92lX8bIbEne0GxLntaOHxrQEVghSWCAc4f53l/LUp6sAOLxnW/70o2F0aZ25p020LIpkuCFxJ2QiXuciMWNMy2GFIEUt37SLq18oYsmGMtI8wrXH9+WKY3rh9Tjf8lWV4FdBKj+oJH1YuhMS19l+XYxpiewvO8WoOreQvPOtxVSHovRol8XD5w9nWPfWe9pEtrkhcavDpBWkkT7SQuKMacmsEKSQbRVBrn9lPu8v3gjAOSO6cfsZA8mJiYoILnZD4ryQdUoW/uEWEmdMS2eFIEV8tnwLv36piI1lAXIz0rjrrMGcPrTLnvl7QuIO8uLr44bEtbIhocakAisELVwwHOWP7y/lyZkrUYWRPdrw0PnD6NYmC3BD4j6tJrIlQvbZbkjcRAuJMyaVWCFowVZuLueXLxaxYN1OvB7hmvF9+MW4XqR5nW/64XVhKt6sILo5in+Q30LijElR9mffAqkqU+aWcuvURVSFInRrk8nD5w/jkB5tnfkhpWpGFYEv3JC487Lx97WQOGNSlRWCFmZnZYgbX5vPtAXfAnDGsC7ceeYgWsXERGhICS4I4h/hJ2t8FpJuJ4ONSWVJLQQichLwME6k9d9U9d4a8y8Crncny4ErVbU4mX1qyT5fuZVr/13Ehp3V5KSnceeZAzlreDcAtFqpnlNNxpEZ34XEZdrJYGNMEguBe+P7x4DjgVJgjohMVdXFMc1WAceo6nYRmQA8CRyWrD61VKFIlIc+KOHxGStQhWHdW/PI+cPJb+ecEA6WBKmcVomWK2nd05yQOCsCxhhXMvcIRgHLVXUlgIi8CJwB7CkEqjorpv3nQLck9qdF+mZrBde8WETx2h14BK46tjfXjO+Dz+shWhGl8r1KQotCeDt6yfpRFmld7GigMWZvyfxU6AqsjZkupe5v+z8D3ok3Q0QuBy4HyM/Pr6/+NWuqyqtfruOWNxZSEYzQJS+Dh84fzqjCtnva7AmJOyaDjCMtJM4YE18yC0HCt7gUkXE4heCoePNV9Umcw0aMHDky5W+TWVYd4nevLWRq8XoAThnSmbvPHExels9C4owxByyZhaAU6B4z3Q1YX7ORiAwB/gZMUNWtSexPizB39TZ++WIR63ZUkeX3ctvpAzn3EOeIWmBewAmJG24hccaYxCXzk2IO0EdECoF1wPnAhbENRCQfeBX4saqWJLEvzV44EuXR6ct5dPoyogpDuuXx8PnDKWyfTWSrGxL3TZi0wjTSD7WQOGNM4pJWCFQ1LCJXAe/hDB99WlUXicgV7vwngFuAdsDjbrBZWFVHJqtPzdXabZX86t9FzPtmOyJw5dheXHtcX/xpnj0hceIVsk7Lwj/UQuKMMQdGVJvXIfeRI0fq3LlzG7sbDeaNonX87rWF7AqE6dQqgz+dN5TRvdrvCYmLbItQ9VGVExKXa0NCjTHxici82r5o20HkJqo8EOaWNxby6pfrADhx4EHce/YQWqf7qJpR5YTEnZONt62XnHMsJM4Y8/1ZIWiCvlqznV++WMSabZVk+DzccupALhjVnci6CGVvlhHdEsU/2ELijDH1wz5GmpBIVPnLjOU8+MEyIlFlQOdWPHLBcHq1zqbqfTckrpWQc0EOvt6+/a/QGGMSYIWgiVi/o4pf/buI/67aBsBlYwr57Yn9SE/zEq2MElwUJH1kOpnHZlpInDGmXlkhaALenr+BG1+dT1l1mA656fzx3KEcld+OwKwAelRMSFyGnQw2xtQ/KwSNqCIQ5vY3F/HS3FIAxvfvyP0Th5BbKpT9pQytUNJ6pOHr4bMiYIxJGisEjWR+6Q5++WIRq7ZUkJ7m4eZTDuaiQd2pereKiiUhvAd5yTrfrg42xiSffco0sGhUefKTlTzw3lLCUaXfQbk8csFw+nXKZdezuwivD5MxNoOM0RYSZ4xpGFYIGtC3O6v59UtFzFrhRCpdMrqA60b3JTPX+TFknpiJpAneDhYSZ4xpOFYIGsh7i77l+lfms6MyRPscP5PPGcoR5a2oeqocLCTOGNOI7JMnyaqCEe58ezHPf7EGgGP6duCB8YPJ+ChM1doq0nqmkT7KQuKMMY3HCkESLVq/k2te+IoVmyvwez3cMKE/F7buQtW/Kon4hKzTs/APsZA4Y0zjskKQBNGo8vRnq7j/3aUEI1H6dMzh4fOGMaBrHpFtEcL9fU5IXI4NCTXGND4rBPVs065qfvNSMZ8s2wLAT0bl85tWPfHMUnSiOiFxZ1tInDGm6bBCUI+mf72RSVPms7UiSJssH4+NG8bABelEFgXxDrGQOGNM02QfS/WgOhThnmlLeG72NwAc27MD93fqj3dGBPIg58IcfL0sJM4Y0zRZIfiBln67i2te+IqlG3fh8wqTTuzHTw8poPzJXfgPTSdznIXEGWOaNisE35Oq8vfZ33DXtCUEw1GGtGvFo/0HkX9Ua8Qj5F2Zh2RYATDGNH1WCL6H6lCEG19dwGtfOXcPu6VfX87a3gGKlfCQML4ePisCxphmwwrBAdpUVs3l/5hH0doddPdn8Ez+UNqVevF28pB1YRZpney/1BjTvNin1gGYX7qDy/8+j2/LqunaOpOX2w7Hvwkyj80k/Yh0xGN7AcaY5scKQYLeLF7Pb6cU0zbiY0x+ex68eBity71OSFx7C4kzxjRfVgj2IxpVHvyghD9PX855aZ24NruArMIMcnPSwa4LM8a0AFYI6lARCPPrl4pYuriMpzMGMczbirSeaWQdkdHYXTPGmHpjhaAWa7dVctnf59Jjs59/Zw7Fky5kTcjCP9hC4owxLYsVgjj+u2obV/5znhMV0bYNmu+l3amtLCTOGNMiWSGoYcrna1j3Thk3Ucgbfbbx6IXDycu0eAhjTMtlhcBVHgjz3JQVjF6Ry3FpXVnUrpKnfnwIPr+NCDLGtGxWCIBZi7ew4rVtXBjtyDqqmTOykhMmdG3sbhljTINI6UKwszLE3dOW8M7cDUzJHMq7mVsZcUFHTujeqrG7ZowxDSZlC8H78zbw9bRtvFyxFq/Xw0dHlPOz8T3xee2EsDEmtSS1EIjIScDDgBf4m6reW2O+uPNPBiqBS1T1y2T2aePOKqa8sJqTNrVlKJ1Y3yXIz87vTe+Oucl8W2OMabKSVghExAs8BhwPlAJzRGSqqi6OaTYB6OM+DgP+4v5b71SV1z8pxTMjyEXSia+1nIVHBLjruGF4LCPIGJPCknkcZBSwXFVXqmoQeBE4o0abM4C/q+NzoLWIdE5GZ+54azF506OMIo8387bQ4+p2nHlCvhUBY0zKS2Yh6AqsjZkudV870DaIyOUiMldE5m7evPl7daZ7myz+GFnN/HFBfnx1b7q0zfpe6zHGmJYmmecI4n3V1u/RBlV9EngSYOTIkfvMT8RPjypkwuBOdM7L/D6LG2NMi5XMPYJSoHvMdDdg/fdoU2+sCBhjzL6SWQjmAH1EpFBE/MD5wNQabaYCF4vjcGCnqm5IYp+MMcbUkLRDQ6oaFpGrgPdwho8+raqLROQKd/4TwDScoaPLcYaPXpqs/hhjjIkvqdcRqOo0nA/72NeeiHmuwC+S2QdjjDF1s8tojTEmxVkhMMaYFGeFwBhjUpwVAmOMSXHinK9tPkRkM/DN91y8PbClHrvTHNg2pwbb5tTwQ7a5h6p2iDej2RWCH0JE5qrqyMbuR0OybU4Nts2pIVnbbIeGjDEmxVkhMMaYFJdqheDJxu5AI7BtTg22zakhKducUucIjDHG7CvV9giMMcbUYIXAGGNSXIssBCJykogsFZHlInJDnPkiIo+48+eLyIjG6Gd9SmCbL3K3db6IzBKRoY3Rz/q0v22OaXeoiEREZGJD9i8ZEtlmERkrIkUiskhEPm7oPta3BH6380TkTREpdre5WacYi8jTIrJJRBbWMr/+P79UtUU9cCKvVwA9AT9QDAyo0eZk4B2cO6QdDnzR2P1ugG0eDbRxn09IhW2OaTcdJwV3YmP3uwF+zq2BxUC+O92xsfvdANt8E3Cf+7wDsA3wN3bff8A2Hw2MABbWMr/eP79a4h7BKGC5qq5U1SDwInBGjTZnAH9Xx+dAaxHp3NAdrUf73WZVnaWq293Jz3HuBtecJfJzBrgaeAXY1JCdS5JEtvlC4FVVXQOgqs19uxPZZgVyRUSAHJxCEG7YbtYfVZ2Jsw21qffPr5ZYCLoCa2OmS93XDrRNc3Kg2/MznG8Uzdl+t1lEugJnAU/QMiTyc+4LtBGRGSIyT0QubrDeJUci2/xn4GCc29wuAH6pqtGG6V6jqPfPr6TemKaRSJzXao6RTaRNc5Lw9ojIOJxCcFRSe5R8iWzzQ8D1qhpxviw2e4lscxpwCDAeyARmi8jnqlqS7M4lSSLbfCJQBBwL9ALeF5FPVLUsyX1rLPX++dUSC0Ep0D1muhvON4UDbdOcJLQ9IjIE+BswQVW3NlDfkiWRbR4JvOgWgfbAySISVtXXG6SH9S/R3+0tqloBVIjITGAo0FwLQSLbfClwrzoH0JeLyCqgP/Dfhulig6v3z6+WeGhoDtBHRApFxA+cD0yt0WYqcLF79v1wYKeqbmjojtaj/W6ziOQDrwI/bsbfDmPtd5tVtVBVC1S1AHgZ+H/NuAhAYr/bbwBjRCRNRLKAw4AlDdzP+pTINq/B2QNCRA4C+gErG7SXDaveP79a3B6BqoZF5CrgPZwRB0+r6iIRucKd/wTOCJKTgeVAJc43imYrwW2+BWgHPO5+Qw5rM05uTHCbW5REtllVl4jIu8B8IAr8TVXjDkNsDhL8Od8JPCsiC3AOm1yvqs02nlpEXgDGAu1FpBS4FfBB8j6/LGLCGGNSXEs8NGSMMeYAWCEwxpgUZ4XAGGNSnBUCY4xJcVYIjDEmxVkhME2WmxhaFPMoqKNteQN2rVYi0kVEXnafDxORk2PmnV5XSmoS+lIgIhc21PuZ5suGj5omS0TKVTWnvts2FBG5BBipqlcl8T3SVDVuwJqIjAV+q6qnJuv9TctgewSm2RCRHBH5UES+FJEFIrJP2qiIdBaRme4exEIRGeO+foKIzHaXnSIi+xQNN6jtIXHu17BQREa5r7cVkdfd7PfP3agOROSYmL2Vr0Qk1/0WvtC9CvYO4Dx3/nkicomI/Fmc/PzVIuJx15MlImtFxCcivUTkXTcw7hMR6R+nn7eJyJMi8h/g7+57fuJu25ciMtptei/OVcZFInKtiHhFZLKIzHG35ef19KMxzV1jZ2/bwx61PYAITphYEfAazpXwrdx57XGurNy9V1vu/vsb4Gb3uRfIddvOBLLd168HbonzfjOAv7rPj8bNgwceBW51nx8LFLnP3wSOdJ/nuP0riFnuEuDPMevfM40TBTHOfX4ezhXAAB8CfdznhwHT4/TzNmAekOlOZwEZ7vM+wFz3+VjgrZjlLgd+5z5PB+YChY39c7ZH4z9aXMSEaVGqVHXY7gkR8QF3i8jROPEJXYGDgG9jlpkDPO22fV1Vi0TkGGAA8Jkbr+EHZtfyni+AkwkvIq1EpDVOUus57uvTRaSdiOQBnwF/EpF/4dwDoFQSTzn9N04B+AgnP+dxdy9lNDAlZj3ptSw/VVWr3Oc+4M8iMgynePatZZkTgCHy3Z3a8nAKx6pEO21aJisEpjm5COcOVIeoakhEVgMZsQ3cD/CjgVOAf4jIZGA78L6qXpDAe9Q8aabUEvurqveKyNs4uS+fi8hxQHWC2zIVuEdE2uLERk8HsoEdscWvDhUxz68FNuKkjHrq6IMAV6vqewn20aQIO0dgmpM8YJNbBMYBPWo2EJEebpu/Ak/h3PLvc+BIEenttskSkdq+NZ/ntjkKJ9VxJ85hpYvc18fixDyXiUgvVV2gqvfhHGapeTx/F86hqX2oajlOTPLDOIdvIurk568SkXPd9xJJ7N7SecAGdW7G8mOcQ2Lx3v894Ep3bwkR6Ssi2Qms37RwtkdgmpN/AW+KyFyc8wZfx2kzFpgkIiGgHLhYVTe7I3heEJHdh1p+R/yM/u0iMgtoBfzUfe024BkRmY+T9vgT9/VfuQUpgnOf4HeA2FsGfgTcICJFwD1x3uvfwBS3z7tdBPxFRH6Hc8jnRZz79NblceAVt4B8xHd7C/OBsIgUA8/iFJ0C4Etxjj1tBs7cz7pNCrDho8a4RGQGznDLuY3dF2Makh0aMsaYFGd7BMYYk+Jsj8AYY1KcFQJjjElxVgiMMSbFWSEwxpgUZ4XAGGNS3P8HlgDysN1n34wAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "plt.plot(fprbnb, tprbnb,lw=2)\n",
    "\n",
    "plt.plot([0,1],[0,1],c='violet',ls='--')\n",
    "plt.xlim([-0.05,1.05])\n",
    "plt.ylim([-0.05,1.05])\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC Score', fontweight=\"bold\", fontsize=20);\n",
    "plt.legend(['BNB'])\n",
    "print(\"BNB ROC AUC score = \", roc_auc_score(y_test, bnb.predict_proba(x_test_pre)[:,1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Hops"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['title',\n",
       " 'bstyle',\n",
       " 'og',\n",
       " 'fg',\n",
       " 'abv',\n",
       " 'ibu',\n",
       " 'srm',\n",
       " 'Mash pH',\n",
       " 'hops',\n",
       " 'others',\n",
       " 'water',\n",
       " 'notes',\n",
       " 'IBU (rager)',\n",
       " 'SRM (daniels)',\n",
       " 'SRM (ebcmorey)',\n",
       " 'SRM (mosher)',\n",
       " 'ABV (alternate)',\n",
       " 'yeast',\n",
       " 'co2',\n",
       " 'chal',\n",
       " 'citra',\n",
       " 'warrior',\n",
       " 'casc',\n",
       " 'saaz',\n",
       " 'azacca',\n",
       " 'colum',\n",
       " 'mag',\n",
       " 'perle',\n",
       " 'simcoe',\n",
       " 'amar',\n",
       " 'cashmere',\n",
       " 'comet',\n",
       " 'galaxy',\n",
       " 'apollo',\n",
       " 'willam',\n",
       " 'goldings',\n",
       " 'heller',\n",
       " 'fug',\n",
       " 'nugg',\n",
       " 'hopshot',\n",
       " 'tett',\n",
       " 'mosaic',\n",
       " 'brewerg',\n",
       " 'northb',\n",
       " 'cent',\n",
       " 'chinook',\n",
       " 'crystal',\n",
       " 'magnum',\n",
       " 'ctz',\n",
       " 'styr',\n",
       " 'idaho',\n",
       " 'nelson',\n",
       " 'exper',\n",
       " 'hall',\n",
       " 'noble',\n",
       " 'dryh',\n",
       " 'lact',\n",
       " 'clone',\n",
       " 'juicy',\n",
       " 'dank',\n",
       " 'wedding',\n",
       " 'whisk',\n",
       " 'trop',\n",
       " 'award',\n",
       " 'best',\n",
       " 'smoke',\n",
       " 'roast',\n",
       " 'thick',\n",
       " 'pine',\n",
       " 'sticky',\n",
       " 'foam',\n",
       " 'hoppy',\n",
       " 'choco',\n",
       " 'coffee',\n",
       " 'straw',\n",
       " 'bisc',\n",
       " 'dark',\n",
       " 'heavy',\n",
       " 'stfruit',\n",
       " 'hazy',\n",
       " 'pheno',\n",
       " 'ester',\n",
       " 'banana',\n",
       " 'bitter',\n",
       " 'clove',\n",
       " 'sour',\n",
       " 'wylon',\n",
       " 'kveik',\n",
       " 'omega',\n",
       " 'us5',\n",
       " 'us4',\n",
       " 'chico',\n",
       " 'yeast_dark',\n",
       " 'wy1056',\n",
       " 'wy1272',\n",
       " 'wlp080',\n",
       " 'wlp80',\n",
       " 'gy054',\n",
       " 'vermont',\n",
       " 'ylager',\n",
       " 'ykolsch',\n",
       " 'ipa',\n",
       " 'saison',\n",
       " 'lager',\n",
       " 'stout',\n",
       " 'kolsch',\n",
       " 'wit',\n",
       " 'puree',\n",
       " 'lovehop',\n",
       " 'strong',\n",
       " 'dry',\n",
       " 'sweet',\n",
       " 'simsty',\n",
       " 'threesty']"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(df.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 804,
   "metadata": {},
   "outputs": [],
   "source": [
    "hops = df[['ipa','chal','citra',\n",
    " 'warrior',\n",
    " 'casc',\n",
    " 'saaz',\n",
    "\n",
    " 'azacca',\n",
    " 'colum',\n",
    " 'mag',\n",
    " 'perle',\n",
    " 'simcoe',\n",
    " 'amar',\n",
    " 'cashmere',\n",
    " 'comet',\n",
    " 'galaxy',\n",
    " 'apollo',\n",
    " 'willam',\n",
    " 'goldings',\n",
    " 'heller',\n",
    " 'fug',\n",
    " 'nugg',\n",
    " 'hopshot',\n",
    " 'tett',\n",
    " 'mosaic',\n",
    " 'brewerg',\n",
    " 'northb',\n",
    " 'cent',\n",
    " 'chinook',\n",
    " 'crystal',\n",
    " 'ctz','styr','idaho','nelson','exper','hall','noble','dryh', 'wylon', 'us5', 'us4', 'chico', 'vermont']].dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 805,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7208991814231157"
      ]
     },
     "execution_count": 805,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.bugu.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 806,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = hops.drop(columns =['ipa'])\n",
    "y= hops['ipa']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 807,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, x_test, y_train, y_test = train_test_split(X, y, test_size=0.2)\n",
    "X_train, x_val, y_train, y_val = train_test_split(X_train, y_train, test_size=0.25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 808,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_pre = preprocessing.scale(X)\n",
    "X_train_pre = preprocessing.scale(X_train)\n",
    "x_test_pre = preprocessing.scale(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 809,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(n_neighbors=30)"
      ]
     },
     "execution_count": 809,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn = KNeighborsClassifier(n_neighbors = 30)\n",
    "knn.fit(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 810,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7999599519423308"
      ]
     },
     "execution_count": 810,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.score(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 811,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7855855855855856"
      ]
     },
     "execution_count": 811,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.score(x_test_pre,y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 799,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, ..., 0, 0, 0])"
      ]
     },
     "execution_count": 799,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.predict(x_test_pre)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 800,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KNN ROC AUC score =  0.8203688758689018\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "plt.plot(fpr, tpr,lw=2)\n",
    "\n",
    "plt.plot([0,1],[0,1],c='violet',ls='--')\n",
    "plt.xlim([-0.05,1.05])\n",
    "plt.ylim([-0.05,1.05])\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC Score', fontweight=\"bold\", fontsize=20);\n",
    "plt.legend(['KNN'])\n",
    "print(\"KNN ROC AUC score = \", roc_auc_score(y_test, knn.predict_proba(x_test_pre)[:,1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 812,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/knnh.pickle', 'wb') as to_write:\n",
    "    pickle.dump(knn, to_write)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Naive Bayes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 813,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7883883883883884"
      ]
     },
     "execution_count": 813,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bnbh = BernoulliNB()\n",
    "bnbh.fit(X_train_pre, y_train)\n",
    "bnbh.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 814,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Expected input with 46 features, got 41 instead",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-814-69a7b6777eed>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mfprbnb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtprbnb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mthresholdsbnb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mroc_curve\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbnb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test_pre\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36mpredict_proba\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m    116\u001b[0m             \u001b[0morder\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mthey\u001b[0m \u001b[0mappear\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mattribute\u001b[0m \u001b[0;34m:\u001b[0m\u001b[0mterm\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0mclasses_\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    117\u001b[0m         \"\"\"\n\u001b[0;32m--> 118\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_log_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    119\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    120\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36mpredict_log_proba\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m     96\u001b[0m         \u001b[0mcheck_is_fitted\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     97\u001b[0m         \u001b[0mX\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_check_X\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 98\u001b[0;31m         \u001b[0mjll\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_joint_log_likelihood\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     99\u001b[0m         \u001b[0;31m# normalize by P(x) = P(f_1, ..., f_n)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    100\u001b[0m         \u001b[0mlog_prob_x\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlogsumexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mjll\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36m_joint_log_likelihood\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m   1009\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1010\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mn_features_X\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mn_features\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1011\u001b[0;31m             raise ValueError(\"Expected input with %d features, got %d instead\"\n\u001b[0m\u001b[1;32m   1012\u001b[0m                              % (n_features, n_features_X))\n\u001b[1;32m   1013\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: Expected input with 46 features, got 41 instead"
     ]
    }
   ],
   "source": [
    "fprbnb, tprbnb, thresholdsbnb = roc_curve(y_test, bnb.predict_proba(x_test_pre)[:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 815,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Expected input with 46 features, got 41 instead",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-815-d0f9f32d3075>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      8\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtitle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'ROC Score'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfontweight\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"bold\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfontsize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m20\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m;\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlegend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'BNB'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 10\u001b[0;31m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"BNB ROC AUC score = \"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mroc_auc_score\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbnb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test_pre\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36mpredict_proba\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m    116\u001b[0m             \u001b[0morder\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mthey\u001b[0m \u001b[0mappear\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mattribute\u001b[0m \u001b[0;34m:\u001b[0m\u001b[0mterm\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0mclasses_\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    117\u001b[0m         \"\"\"\n\u001b[0;32m--> 118\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_log_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    119\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    120\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36mpredict_log_proba\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m     96\u001b[0m         \u001b[0mcheck_is_fitted\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     97\u001b[0m         \u001b[0mX\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_check_X\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 98\u001b[0;31m         \u001b[0mjll\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_joint_log_likelihood\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     99\u001b[0m         \u001b[0;31m# normalize by P(x) = P(f_1, ..., f_n)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    100\u001b[0m         \u001b[0mlog_prob_x\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlogsumexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mjll\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36m_joint_log_likelihood\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m   1009\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1010\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mn_features_X\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mn_features\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1011\u001b[0;31m             raise ValueError(\"Expected input with %d features, got %d instead\"\n\u001b[0m\u001b[1;32m   1012\u001b[0m                              % (n_features, n_features_X))\n\u001b[1;32m   1013\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: Expected input with 46 features, got 41 instead"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "plt.plot(fprbnb, tprbnb,lw=2)\n",
    "\n",
    "plt.plot([0,1],[0,1],c='violet',ls='--')\n",
    "plt.xlim([-0.05,1.05])\n",
    "plt.ylim([-0.05,1.05])\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC Score', fontweight=\"bold\", fontsize=20);\n",
    "plt.legend(['BNB'])\n",
    "print(\"BNB ROC AUC score = \", roc_auc_score(y_test, bnb.predict_proba(x_test_pre)[:,1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 816,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/bngbh.pickle', 'wb') as to_write:\n",
    "    pickle.dump(bnbh, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 817,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 818,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier()"
      ]
     },
     "execution_count": 818,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest = RandomForestClassifier()\n",
    "forest.fit(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 819,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8460152182619143"
      ]
     },
     "execution_count": 819,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest.score(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 820,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7821821821821822"
      ]
     },
     "execution_count": 820,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 821,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/foresth.pickle', 'wb') as to_write:\n",
    "    pickle.dump(forest, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 822,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.svm import SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 823,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SVC(probability=True)"
      ]
     },
     "execution_count": 823,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svc = SVC(probability = True)\n",
    "svc.fit(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 824,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.795995995995996"
      ]
     },
     "execution_count": 824,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svc.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 825,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/svch.pickle', 'wb') as to_write:\n",
    "    pickle.dump(svc, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 826,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7964891202776665"
      ]
     },
     "execution_count": 826,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logith = LogisticRegression()\n",
    "logith.fit(X_train_pre,y_train)\n",
    "logith.score(X_train_pre,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 827,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.7933933933933934,\n",
       " array([[-0.00668492,  0.18563875,  0.07698282,  0.03086803, -0.18426905,\n",
       "          0.06680304,  0.1519236 ,  0.04635977, -0.08714037,  0.17102689,\n",
       "          0.08439475,  0.02737368,  0.03659551,  0.09059987,  0.03907532,\n",
       "         -0.02393347, -0.07636112,  0.03155504, -0.09620349,  0.06458944,\n",
       "         -0.00322434, -0.10605241,  0.0196128 , -0.0017998 , -0.03299172,\n",
       "          0.12026747,  0.08418346, -0.00190188,  0.01490764, -0.02972287,\n",
       "         -0.0012219 ,  0.04268018,  0.03012307,  0.        , -0.05529874,\n",
       "          0.61392254,  0.04724673,  0.04723347,  0.0139961 , -0.00406359,\n",
       "          0.03822783]]))"
      ]
     },
     "execution_count": 827,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logith.score(x_test_pre,y_test), logith.coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 828,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/logith.pickle', 'wb') as to_write:\n",
    "    pickle.dump(logith, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 829,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7935935935935936"
      ]
     },
     "execution_count": 829,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbh = xgb.XGBClassifier()\n",
    "xgbh.fit(X_train, y_train)\n",
    "xgbh.score(x_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 830,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/xgbh.pickle', 'wb') as to_write:\n",
    "    pickle.dump(xgbh, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 399,
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_test, knn.predict_proba(x_test_pre)[:,1])\n",
    "fprlog, tprlog, thresholdslog = roc_curve(y_test, logit.predict_proba(x_test_pre)[:,1])\n",
    "fprnb, tprnb, thresholdsnb = roc_curve(y_test, bnb.predict_proba(x_test_pre)[:,1])\n",
    "fprrf, tprrf, thresholdsrf = roc_curve(y_test, forest.predict_proba(x_test_pre)[:,1])\n",
    "\n",
    "fprsvc, tprsvc, thresholdssvc = roc_curve(y_test, svc.predict_proba(x_test_pre)[:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 400,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KNN ROC AUC score =  0.8751513175364334\n",
      "Logit ROC AUC score =  0.898512374287901\n",
      "Naive Bayes ROC AUC score =  0.8968837234211097\n",
      "Random Forest ROC AUC score =  0.9339308586245137\n",
      "SVC ROC AUC score =  0.9260569076664575\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (15,10))\n",
    "plt.plot(fpr, tpr,lw=2)\n",
    "plt.plot(fprlog, tprlog,lw=2)\n",
    "plt.plot(fprnb, tprnb,lw=2)\n",
    "plt.plot(fprrf, tprrf,lw=2)\n",
    "\n",
    "plt.plot(fprsvc, tprsvc,lw=2)\n",
    "plt.plot([0,1],[0,1],c='violet',ls='--')\n",
    "plt.xlim([-0.05,1.05])\n",
    "plt.ylim([-0.05,1.05])\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC Scores', fontweight=\"bold\", fontsize=20);\n",
    "plt.legend(['KNN','Logit', 'Bernoulli NB', 'Random Forest','SVC'])\n",
    "print(\"KNN ROC AUC score = \", roc_auc_score(y_test, knn.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"Logit ROC AUC score = \", roc_auc_score(y_test, logit.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"Naive Bayes ROC AUC score = \", roc_auc_score(y_test, bnb.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"Random Forest ROC AUC score = \", roc_auc_score(y_test, forest.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"SVC ROC AUC score = \", roc_auc_score(y_test, svc.predict_proba(x_test_pre)[:,1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Hops OverSample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['title',\n",
       " 'bstyle',\n",
       " 'og',\n",
       " 'fg',\n",
       " 'abv',\n",
       " 'ibu',\n",
       " 'srm',\n",
       " 'Mash pH',\n",
       " 'hops',\n",
       " 'others',\n",
       " 'water',\n",
       " 'notes',\n",
       " 'IBU (rager)',\n",
       " 'SRM (daniels)',\n",
       " 'SRM (ebcmorey)',\n",
       " 'SRM (mosher)',\n",
       " 'ABV (alternate)',\n",
       " 'yeast',\n",
       " 'co2',\n",
       " 'chal',\n",
       " 'citra',\n",
       " 'warrior',\n",
       " 'casc',\n",
       " 'saaz',\n",
       " 'azacca',\n",
       " 'colum',\n",
       " 'mag',\n",
       " 'perle',\n",
       " 'simcoe',\n",
       " 'amar',\n",
       " 'cashmere',\n",
       " 'comet',\n",
       " 'galaxy',\n",
       " 'apollo',\n",
       " 'willam',\n",
       " 'goldings',\n",
       " 'heller',\n",
       " 'fug',\n",
       " 'nugg',\n",
       " 'hopshot',\n",
       " 'tett',\n",
       " 'mosaic',\n",
       " 'brewerg',\n",
       " 'northb',\n",
       " 'cent',\n",
       " 'chinook',\n",
       " 'crystal',\n",
       " 'magnum',\n",
       " 'ctz',\n",
       " 'styr',\n",
       " 'idaho',\n",
       " 'nelson',\n",
       " 'exper',\n",
       " 'hall',\n",
       " 'noble',\n",
       " 'dryh',\n",
       " 'lact',\n",
       " 'clone',\n",
       " 'juicy',\n",
       " 'dank',\n",
       " 'wedding',\n",
       " 'whisk',\n",
       " 'trop',\n",
       " 'award',\n",
       " 'best',\n",
       " 'smoke',\n",
       " 'roast',\n",
       " 'thick',\n",
       " 'pine',\n",
       " 'sticky',\n",
       " 'foam',\n",
       " 'hoppy',\n",
       " 'choco',\n",
       " 'coffee',\n",
       " 'straw',\n",
       " 'bisc',\n",
       " 'dark',\n",
       " 'heavy',\n",
       " 'stfruit',\n",
       " 'hazy',\n",
       " 'pheno',\n",
       " 'ester',\n",
       " 'banana',\n",
       " 'bitter',\n",
       " 'clove',\n",
       " 'sour',\n",
       " 'wylon',\n",
       " 'kveik',\n",
       " 'omega',\n",
       " 'us5',\n",
       " 'us4',\n",
       " 'chico',\n",
       " 'yeast_dark',\n",
       " 'wy1056',\n",
       " 'wy1272',\n",
       " 'wlp080',\n",
       " 'wlp80',\n",
       " 'gy054',\n",
       " 'vermont',\n",
       " 'ylager',\n",
       " 'ykolsch',\n",
       " 'ipa',\n",
       " 'saison',\n",
       " 'lager',\n",
       " 'stout',\n",
       " 'kolsch',\n",
       " 'wit',\n",
       " 'puree',\n",
       " 'lovehop',\n",
       " 'strong',\n",
       " 'dry',\n",
       " 'sweet',\n",
       " 'simsty',\n",
       " 'threesty']"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(df.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 904,
   "metadata": {},
   "outputs": [],
   "source": [
    "ohops = df[['ipa','chal','citra',\n",
    " 'warrior',\n",
    " 'casc',\n",
    " 'saaz',\n",
    "\n",
    " 'azacca',\n",
    " 'colum',\n",
    " 'mag',\n",
    " 'perle',\n",
    " 'simcoe',\n",
    " 'amar',\n",
    " 'cashmere',\n",
    " 'comet',\n",
    " 'galaxy',\n",
    " 'apollo',\n",
    " 'willam',\n",
    " 'goldings',\n",
    " 'heller',\n",
    " 'fug',\n",
    " 'nugg',\n",
    " 'hopshot',\n",
    " 'tett',\n",
    " 'mosaic',\n",
    " 'brewerg',\n",
    " 'northb',\n",
    " 'cent',\n",
    " 'chinook',\n",
    " 'crystal',\n",
    " 'ctz','styr','idaho','nelson','exper','hall','noble','dryh', 'wylon', 'us5', 'us4', 'chico', 'vermont']].dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 805,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7208991814231157"
      ]
     },
     "execution_count": 805,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.bugu.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 906,
   "metadata": {},
   "outputs": [],
   "source": [
    "Xo = ohops.drop(columns =['ipa'])\n",
    "yo= ohops['ipa']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 908,
   "metadata": {},
   "outputs": [],
   "source": [
    "Xo_train, xo_test, yo_train, yo_test = train_test_split(Xo, yo, test_size=0.2)\n",
    "Xo_train, xo_val, yo_train, yo_val = train_test_split(Xo_train, yo_train, test_size=0.25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 935,
   "metadata": {},
   "outputs": [],
   "source": [
    "from imblearn.over_sampling import SMOTE, ADASYN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 936,
   "metadata": {},
   "outputs": [],
   "source": [
    "ros = ADASYN()\n",
    "Xo_resampled, yo_resampled = ros.fit_sample(Xo_train,yo_train)\n",
    "#X_pre = preprocessing.scale(X)\n",
    "Xo_train_pre = preprocessing.scale(Xo_resampled)\n",
    "#x_test_pre = preprocessing.scale(x_test_resampled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 937,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_test_pre = preprocessing.scale(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 938,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(n_neighbors=30)"
      ]
     },
     "execution_count": 938,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn = KNeighborsClassifier(n_neighbors = 30)\n",
    "knn.fit(Xo_train_pre, yo_resampled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 939,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6438800674944837"
      ]
     },
     "execution_count": 939,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.score(Xo_train_pre, yo_resampled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 940,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7793793793793794"
      ]
     },
     "execution_count": 940,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.score(x_test_pre,y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 799,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, ..., 0, 0, 0])"
      ]
     },
     "execution_count": 799,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn.predict(x_test_pre)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Need to beat .82, doesn't"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 941,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KNN ROC AUC score =  0.7848085013881652\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "plt.plot(fpr, tpr,lw=2)\n",
    "\n",
    "plt.plot([0,1],[0,1],c='violet',ls='--')\n",
    "plt.xlim([-0.05,1.05])\n",
    "plt.ylim([-0.05,1.05])\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC Score', fontweight=\"bold\", fontsize=20);\n",
    "plt.legend(['KNN'])\n",
    "print(\"KNN ROC AUC score = \", roc_auc_score(y_test, knn.predict_proba(x_test_pre)[:,1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 812,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/knnh.pickle', 'wb') as to_write:\n",
    "    pickle.dump(knn, to_write)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Naive Bayes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 918,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7561561561561562"
      ]
     },
     "execution_count": 918,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bnbh = BernoulliNB()\n",
    "bnbh.fit(Xo_train_pre, yo_resampled)\n",
    "bnbh.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 814,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Expected input with 46 features, got 41 instead",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-814-69a7b6777eed>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mfprbnb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtprbnb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mthresholdsbnb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mroc_curve\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbnb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test_pre\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36mpredict_proba\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m    116\u001b[0m             \u001b[0morder\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mthey\u001b[0m \u001b[0mappear\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mattribute\u001b[0m \u001b[0;34m:\u001b[0m\u001b[0mterm\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0mclasses_\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    117\u001b[0m         \"\"\"\n\u001b[0;32m--> 118\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_log_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    119\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    120\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36mpredict_log_proba\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m     96\u001b[0m         \u001b[0mcheck_is_fitted\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     97\u001b[0m         \u001b[0mX\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_check_X\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 98\u001b[0;31m         \u001b[0mjll\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_joint_log_likelihood\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     99\u001b[0m         \u001b[0;31m# normalize by P(x) = P(f_1, ..., f_n)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    100\u001b[0m         \u001b[0mlog_prob_x\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlogsumexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mjll\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36m_joint_log_likelihood\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m   1009\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1010\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mn_features_X\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mn_features\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1011\u001b[0;31m             raise ValueError(\"Expected input with %d features, got %d instead\"\n\u001b[0m\u001b[1;32m   1012\u001b[0m                              % (n_features, n_features_X))\n\u001b[1;32m   1013\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: Expected input with 46 features, got 41 instead"
     ]
    }
   ],
   "source": [
    "fprbnb, tprbnb, thresholdsbnb = roc_curve(y_test, bnb.predict_proba(x_test_pre)[:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 815,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Expected input with 46 features, got 41 instead",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-815-d0f9f32d3075>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      8\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtitle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'ROC Score'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfontweight\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"bold\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfontsize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m20\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m;\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlegend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'BNB'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 10\u001b[0;31m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"BNB ROC AUC score = \"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mroc_auc_score\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbnb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test_pre\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36mpredict_proba\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m    116\u001b[0m             \u001b[0morder\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mthey\u001b[0m \u001b[0mappear\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mattribute\u001b[0m \u001b[0;34m:\u001b[0m\u001b[0mterm\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0mclasses_\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    117\u001b[0m         \"\"\"\n\u001b[0;32m--> 118\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_log_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    119\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    120\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36mpredict_log_proba\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m     96\u001b[0m         \u001b[0mcheck_is_fitted\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     97\u001b[0m         \u001b[0mX\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_check_X\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 98\u001b[0;31m         \u001b[0mjll\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_joint_log_likelihood\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     99\u001b[0m         \u001b[0;31m# normalize by P(x) = P(f_1, ..., f_n)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    100\u001b[0m         \u001b[0mlog_prob_x\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlogsumexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mjll\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/naive_bayes.py\u001b[0m in \u001b[0;36m_joint_log_likelihood\u001b[0;34m(self, X)\u001b[0m\n\u001b[1;32m   1009\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1010\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mn_features_X\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mn_features\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1011\u001b[0;31m             raise ValueError(\"Expected input with %d features, got %d instead\"\n\u001b[0m\u001b[1;32m   1012\u001b[0m                              % (n_features, n_features_X))\n\u001b[1;32m   1013\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: Expected input with 46 features, got 41 instead"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "plt.plot(fprbnb, tprbnb,lw=2)\n",
    "\n",
    "plt.plot([0,1],[0,1],c='violet',ls='--')\n",
    "plt.xlim([-0.05,1.05])\n",
    "plt.ylim([-0.05,1.05])\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC Score', fontweight=\"bold\", fontsize=20);\n",
    "plt.legend(['BNB'])\n",
    "print(\"BNB ROC AUC score = \", roc_auc_score(y_test, bnb.predict_proba(x_test_pre)[:,1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 816,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/bngbh.pickle', 'wb') as to_write:\n",
    "    pickle.dump(bnbh, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 817,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 919,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier()"
      ]
     },
     "execution_count": 919,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest = RandomForestClassifier()\n",
    "forest.fit(Xo_train_pre, yo_resampled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 920,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7818048324656254"
      ]
     },
     "execution_count": 920,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest.score(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 921,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7847847847847848"
      ]
     },
     "execution_count": 921,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 821,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/foresth.pickle', 'wb') as to_write:\n",
    "    pickle.dump(forest, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 822,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.svm import SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 922,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SVC(probability=True)"
      ]
     },
     "execution_count": 922,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svc = SVC(probability = True)\n",
    "svc.fit(Xo_train_pre, yo_resampled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 923,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7485485485485486"
      ]
     },
     "execution_count": 923,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svc.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 825,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/svch.pickle', 'wb') as to_write:\n",
    "    pickle.dump(svc, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 924,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7363363363363363"
      ]
     },
     "execution_count": 924,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logith = LogisticRegression()\n",
    "logith.fit(Xo_train_pre,yo_resampled)\n",
    "logith.score(x_test_pre,y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 827,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.7933933933933934,\n",
       " array([[-0.00668492,  0.18563875,  0.07698282,  0.03086803, -0.18426905,\n",
       "          0.06680304,  0.1519236 ,  0.04635977, -0.08714037,  0.17102689,\n",
       "          0.08439475,  0.02737368,  0.03659551,  0.09059987,  0.03907532,\n",
       "         -0.02393347, -0.07636112,  0.03155504, -0.09620349,  0.06458944,\n",
       "         -0.00322434, -0.10605241,  0.0196128 , -0.0017998 , -0.03299172,\n",
       "          0.12026747,  0.08418346, -0.00190188,  0.01490764, -0.02972287,\n",
       "         -0.0012219 ,  0.04268018,  0.03012307,  0.        , -0.05529874,\n",
       "          0.61392254,  0.04724673,  0.04723347,  0.0139961 , -0.00406359,\n",
       "          0.03822783]]))"
      ]
     },
     "execution_count": 827,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logith.score(x_test_pre,y_test), logith.coef_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 828,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/logith.pickle', 'wb') as to_write:\n",
    "    pickle.dump(logith, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 925,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7685685685685686"
      ]
     },
     "execution_count": 925,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbh = xgb.XGBClassifier()\n",
    "xgbh.fit(Xo_train_pre, yo_resampled)\n",
    "xgbh.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 830,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/xgbh.pickle', 'wb') as to_write:\n",
    "    pickle.dump(xgbh, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 926,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'logit' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-926-5c9ea6a6e2e5>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mfpr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtpr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mthresholds\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mroc_curve\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mknn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test_pre\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mfprlog\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtprlog\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mthresholdslog\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mroc_curve\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlogit\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test_pre\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0mfprnb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtprnb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mthresholdsnb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mroc_curve\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbnb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test_pre\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mfprrf\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtprrf\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mthresholdsrf\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mroc_curve\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mforest\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test_pre\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'logit' is not defined"
     ]
    }
   ],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_test, knn.predict_proba(x_test_pre)[:,1])\n",
    "fprlog, tprlog, thresholdslog = roc_curve(y_test, logith.predict_proba(x_test_pre)[:,1])\n",
    "fprnb, tprnb, thresholdsnb = roc_curve(y_test, bnb.predict_proba(x_test_pre)[:,1])\n",
    "fprrf, tprrf, thresholdsrf = roc_curve(y_test, forest.predict_proba(x_test_pre)[:,1])\n",
    "\n",
    "fprsvc, tprsvc, thresholdssvc = roc_curve(y_test, svc.predict_proba(x_test_pre)[:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.figure(figsize = (15,10))\n",
    "plt.plot(fpr, tpr,lw=2)\n",
    "plt.plot(fprlog, tprlog,lw=2)\n",
    "plt.plot(fprnb, tprnb,lw=2)\n",
    "plt.plot(fprrf, tprrf,lw=2)\n",
    "\n",
    "plt.plot(fprsvc, tprsvc,lw=2)\n",
    "plt.plot([0,1],[0,1],c='violet',ls='--')\n",
    "plt.xlim([-0.05,1.05])\n",
    "plt.ylim([-0.05,1.05])\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC Scores', fontweight=\"bold\", fontsize=20);\n",
    "plt.legend(['KNN','Logit', 'Bernoulli NB', 'Random Forest','SVC'])\n",
    "print(\"KNN ROC AUC score = \", roc_auc_score(y_test, knn.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"Logit ROC AUC score = \", roc_auc_score(y_test, logit.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"Naive Bayes ROC AUC score = \", roc_auc_score(y_test, bnb.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"Random Forest ROC AUC score = \", roc_auc_score(y_test, forest.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"SVC ROC AUC score = \", roc_auc_score(y_test, svc.predict_proba(x_test_pre)[:,1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {},
   "outputs": [],
   "source": [
    "import word_search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 834,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['title',\n",
       " 'bstyle',\n",
       " 'og',\n",
       " 'fg',\n",
       " 'abv',\n",
       " 'ibu',\n",
       " 'srm',\n",
       " 'Mash pH',\n",
       " 'hops',\n",
       " 'others',\n",
       " 'water',\n",
       " 'notes',\n",
       " 'IBU (rager)',\n",
       " 'SRM (daniels)',\n",
       " 'SRM (ebcmorey)',\n",
       " 'SRM (mosher)',\n",
       " 'ABV (alternate)',\n",
       " 'yeast',\n",
       " 'co2',\n",
       " 'chal',\n",
       " 'citra',\n",
       " 'warrior',\n",
       " 'casc',\n",
       " 'saaz',\n",
       " 'azacca',\n",
       " 'colum',\n",
       " 'mag',\n",
       " 'perle',\n",
       " 'simcoe',\n",
       " 'amar',\n",
       " 'cashmere',\n",
       " 'comet',\n",
       " 'galaxy',\n",
       " 'apollo',\n",
       " 'willam',\n",
       " 'goldings',\n",
       " 'heller',\n",
       " 'fug',\n",
       " 'nugg',\n",
       " 'hopshot',\n",
       " 'tett',\n",
       " 'mosaic',\n",
       " 'brewerg',\n",
       " 'northb',\n",
       " 'cent',\n",
       " 'chinook',\n",
       " 'crystal',\n",
       " 'magnum',\n",
       " 'ctz',\n",
       " 'styr',\n",
       " 'idaho',\n",
       " 'nelson',\n",
       " 'exper',\n",
       " 'hall',\n",
       " 'noble',\n",
       " 'dryh',\n",
       " 'lact',\n",
       " 'hopcmp',\n",
       " 'clone',\n",
       " 'juicy',\n",
       " 'dank',\n",
       " 'wedding',\n",
       " 'whisk',\n",
       " 'trop',\n",
       " 'award',\n",
       " 'best',\n",
       " 'smoke',\n",
       " 'roast',\n",
       " 'thick',\n",
       " 'pine',\n",
       " 'sticky',\n",
       " 'foam',\n",
       " 'hoppy',\n",
       " 'choco',\n",
       " 'coffee',\n",
       " 'straw',\n",
       " 'bisc',\n",
       " 'dark',\n",
       " 'heavy',\n",
       " 'stfruit',\n",
       " 'hazy',\n",
       " 'pheno',\n",
       " 'ester',\n",
       " 'banana',\n",
       " 'bitter',\n",
       " 'clove',\n",
       " 'sour',\n",
       " 'wylon',\n",
       " 'kveik',\n",
       " 'omega',\n",
       " 'us5',\n",
       " 'us4',\n",
       " 'chico',\n",
       " 'yeast_dark',\n",
       " 'wy1056',\n",
       " 'wy1272',\n",
       " 'wlp080',\n",
       " 'wlp80',\n",
       " 'vermont',\n",
       " 'ylager',\n",
       " 'ykolsch',\n",
       " 'ipa',\n",
       " 'saison',\n",
       " 'lager',\n",
       " 'stout',\n",
       " 'kolsch',\n",
       " 'wit',\n",
       " 'puree',\n",
       " 'lovehop',\n",
       " 'strong',\n",
       " 'dry',\n",
       " 'sweet',\n",
       " 'bugu',\n",
       " 'simsty',\n",
       " 'threesty',\n",
       " 'foursty',\n",
       " 'malt',\n",
       " 'agg',\n",
       " 'balance',\n",
       " 'complex',\n",
       " 'crisp',\n",
       " 'smooth',\n",
       " 'diac',\n",
       " 'apple',\n",
       " 'floral',\n",
       " 'fruity',\n",
       " 'robust',\n",
       " 'wood',\n",
       " 'vanilla',\n",
       " 'cherry',\n",
       " 'funk',\n",
       " 'grapef',\n",
       " 'solvent']"
      ]
     },
     "execution_count": 834,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(df.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 836,
   "metadata": {},
   "outputs": [],
   "source": [
    "no_null_notes = df.dropna(subset = ['notes'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 886,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_df = no_null_notes[['threesty', 'dryh', 'lact', 'clone', 'juicy', 'dank', 'wedding', 'whisk', 'trop',\\\n",
    "              'award', 'best', 'smoke', 'roast', 'thick',\\\n",
    "              'heavy', 'stfruit', 'hazy', 'pheno', 'banana','clove', \\\n",
    "              'bitter','puree','lovehop', \\\n",
    "              'strong', 'dry', 'sweet', 'sour', 'dark', 'pine', \\\n",
    "            'sticky', 'foam', 'hoppy','bisc','ester', 'straw', 'choco','coffee','malt',\\\n",
    " 'agg',\n",
    " 'balance',\n",
    " 'complex',\n",
    " 'crisp',\n",
    " 'smooth',\n",
    " 'diac',\n",
    " 'apple',\n",
    " 'floral',\n",
    " 'fruity',\n",
    " 'robust',\n",
    " 'wood',\n",
    " 'vanilla',\n",
    " 'cherry',\n",
    " 'funk',\n",
    " 'grapef',\n",
    " 'solvent' ]].dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 843,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(24972, 16767)"
      ]
     },
     "execution_count": 843,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df), len(word_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 887,
   "metadata": {},
   "outputs": [],
   "source": [
    "X3b = word_df.drop(columns =['threesty'])\n",
    "y3b = word_df['threesty']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 854,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['dryh', 'lact', 'clone', 'juicy', 'dank', 'wedding', 'whisk', 'trop',\n",
       "       'award', 'best', 'smoke', 'roast', 'thick', 'heavy', 'stfruit', 'hazy',\n",
       "       'pheno', 'banana', 'clove', 'bitter', 'puree', 'lovehop', 'strong',\n",
       "       'dry', 'sweet', 'sour', 'dark', 'sticky', 'pine', 'sticky', 'foam',\n",
       "       'hoppy', 'bisc', 'ester', 'straw', 'choco', 'coffee', 'malt', 'agg',\n",
       "       'balance', 'complex', 'crisp', 'smooth', 'diac', 'apple', 'floral',\n",
       "       'fruity', 'robust', 'wood', 'vanilla', 'cherry', 'funk', 'grapef',\n",
       "       'solvent'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 854,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X3b.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 888,
   "metadata": {},
   "outputs": [],
   "source": [
    "X3b_train, x3b_test, y3b_train, y3b_test = train_test_split(X3b, y3b, test_size=0.2)\n",
    "X3b_train, x3b_val, y3b_train, y3b_val = train_test_split(X3b_train, y3b_train, test_size=0.25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 889,
   "metadata": {},
   "outputs": [],
   "source": [
    "X3b_pre = preprocessing.scale(X3b)\n",
    "X3b_train_pre = preprocessing.scale(X3b_train)\n",
    "x3b_test_pre = preprocessing.scale(x3b_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "ros = RandomOverSampler(random_state=0)\n",
    "X_resampled, y_resampled = ros.fit_sample(X3b_train_pre,y3b_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 890,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier()"
      ]
     },
     "execution_count": 890,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest3b = RandomForestClassifier() #, 4:2, 5:2, 8:2, 15:2 , 16: 2, 20:2, 22:2, 23:2\n",
    "forest3b.fit(X3b_train_pre, y3b_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 891,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7248235411074659"
      ]
     },
     "execution_count": 891,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest3b.score(X3b_train_pre, y3b_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 892,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6332737030411449"
      ]
     },
     "execution_count": 892,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest3b.score(x3b_test_pre, y3b_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Strong/Dark Ales'], dtype=object)"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest3b.predict(bro_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 893,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/f3.pickle', 'wb') as to_write:\n",
    "    pickle.dump(forest3b, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.3362931224679302, 0.29885397700135047, 0.3917841227855375, 0.402244950407192, 0.4115505116872223, 0.4534497671283691, 0.4570836434651687, 0.4522929854846178, 0.45895682964852264, 0.4595608896884296, 0.4588451518817627, 0.4580204638521647, 0.46325031700120506, 0.4638004928974759, 0.4680390657010758, 0.4641858039223292, 0.4650108859304803, 0.4676538455943501, 0.46908459386266266, 0.4686448531854076, 0.46787420082965825, 0.46655243309031924, 0.46693792595142797, 0.4680935256595201, 0.46814837959651756, 0.4683135172223178, 0.4689193956247773, 0.4679834238069723, 0.47029568393464566, 0.4705710749432065, 0.47194730264098983, 0.4716169061652188, 0.47304832116646744, 0.47266279799931626, 0.47167185102034387, 0.4726628586114015, 0.4720024596384126, 0.4725527870648961, 0.4726630404476567, 0.47282799623720184, 0.47238752821492574, 0.47178161950642367, 0.47178171042455136, 0.47249741792517563, 0.4717265837331711, 0.47156129457715795, 0.4721121978184499, 0.47348860735248843, 0.4738738880712992, 0.47370868983341374]\n"
     ]
    }
   ],
   "source": [
    "k_options = list(range(1, 51))\n",
    "k_scores = []\n",
    "for k in k_options:\n",
    "    knn_bro = KNeighborsClassifier(n_neighbors=k)\n",
    "    scores = cross_val_score(knn_bro, X, y, cv=10, scoring='accuracy')\n",
    "    k_scores.append(scores.mean())\n",
    "print(k_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7faaac51de50>]"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dict_k = (dict(zip(range(50), k_scores)))\n",
    "plt.plot(k_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 885,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pale Ales and Lagers                   8395\n",
       "Strong/Dark Ales                       6639\n",
       "Other (Sours, Wild, Holiday styles)    1733\n",
       "Name: threesty, dtype: int64"
      ]
     },
     "execution_count": 885,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y3b.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [],
   "source": [
    "max_kv = max(dict_k.values())\n",
    "for key, value in dict_k.items():\n",
    "    if value == max_kv:\n",
    "        best_k = key"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 894,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(n_neighbors=43)"
      ]
     },
     "execution_count": 894,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn3b = KNeighborsClassifier(n_neighbors = 43)\n",
    "knn3b.fit(X3b_train_pre, y3b_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 895,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.635450840043742"
      ]
     },
     "execution_count": 895,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn3b.score(X3b_train, y3b_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 896,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6356589147286822"
      ]
     },
     "execution_count": 896,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn3b.score(x3b_test,y3b_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 897,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/knn3b.pickle', 'wb') as to_write:\n",
    "    pickle.dump(knn3b, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 898,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.6808827915299731, 0.6392367322599881)"
      ]
     },
     "execution_count": 898,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svc3b = SVC()\n",
    "svc3b.fit(X3b_train_pre, y3b_train)\n",
    "svc3b.score(X3b_train_pre, y3b_train), svc3b.score(x3b_test_pre, y3b_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 899,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/svc3b.pickle', 'wb') as to_write:\n",
    "    pickle.dump(svc3b, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 367,
   "metadata": {},
   "outputs": [],
   "source": [
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 900,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6443053070960048"
      ]
     },
     "execution_count": 900,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbm = xgb.XGBClassifier(learning_rate = .02, n_estimators = 100 ,max_depth = 4, min_child_weight = 1, objective='multi:softmax')\n",
    "xgbm.fit(X3b_train_pre, y3b_train)\n",
    "xgbm.score(x3b_test_pre, y3b_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 901,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/xgb3b.pickle', 'wb') as to_write:\n",
    "    pickle.dump(xgbm, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 902,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.6108957152798489, 0.6150864639236733)"
      ]
     },
     "execution_count": 902,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnb3b = MultinomialNB()\n",
    "mnb3b.fit(abs(X3b_train_pre), y3b_train)\n",
    "mnb3b.score(abs(X3b_train_pre), y3b_train), mnb3b.score(abs(x3b_test_pre), y3b_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 903,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/mnb3b.pickle', 'wb') as to_write:\n",
    "    pickle.dump(mnb3b, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 370,
   "metadata": {},
   "outputs": [],
   "source": [
    "abv = 6\n",
    "srm = 6\n",
    "ibu = 75\n",
    "bugu = .72\n",
    "hopcmp = 90\n",
    "nums = [abv, srm, ibu]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 371,
   "metadata": {},
   "outputs": [],
   "source": [
    "review = \"wow, mango and papaya tropical tastes. Juicy and hazy. A real hoppy beer for hop lovers!\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 372,
   "metadata": {},
   "outputs": [],
   "source": [
    "bro = word_search.search_desc(review)\n",
    "    #'This beer is incredibly dank and bitter. Bone dry and delicious. Only for true hop lovers. Hazy stone fruit bomb. Tropical.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 373,
   "metadata": {},
   "outputs": [],
   "source": [
    "w_nums = np.append(bro, nums)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 374,
   "metadata": {},
   "outputs": [],
   "source": [
    "bro_df = pd.DataFrame(w_nums).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 349,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>31</th>\n",
       "      <th>32</th>\n",
       "      <th>33</th>\n",
       "      <th>34</th>\n",
       "      <th>35</th>\n",
       "      <th>36</th>\n",
       "      <th>37</th>\n",
       "      <th>38</th>\n",
       "      <th>39</th>\n",
       "      <th>40</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>0.72</td>\n",
       "      <td>90.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1 rows × 41 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    0    1    2    3    4    5    6    7    8    9   ...   31   32   33   34  \\\n",
       "0  0.0  0.0  0.0  1.0  0.0  0.0  0.0  2.0  0.0  0.0  ...  0.0  0.0  0.0  0.0   \n",
       "\n",
       "    35   36   37    38    39    40  \n",
       "0  0.0  6.0  6.0  75.0  0.72  90.0  \n",
       "\n",
       "[1 rows x 41 columns]"
      ]
     },
     "execution_count": 349,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bro_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 350,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Strong/Dark Ales'], dtype=object)"
      ]
     },
     "execution_count": 350,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svc3b.predict(bro_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 375,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('Strong/Dark Ales',\n",
       " 'Pale Ales and Lagers',\n",
       " 'Pale Ales and Lagers',\n",
       " 'Pale Ales and Lagers')"
      ]
     },
     "execution_count": 375,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(forest3b.predict(bro_df)[0], knn3b.predict(bro_df)[0], svc3b.predict(bro_df)[0], mnb3b.predict(bro_df)[0]) #xgbm.predict(bro_df)[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 376,
   "metadata": {},
   "outputs": [],
   "source": [
    "forpredtabsvc = pd.DataFrame(svc3b.predict(x3b_test_pre))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 377,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pale Ales and Lagers                   1792\n",
       "Strong/Dark Ales                       1179\n",
       "Other (Sours, Wild, Holiday styles)      45\n",
       "dtype: int64"
      ]
     },
     "execution_count": 377,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forpredtabsvc.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 378,
   "metadata": {},
   "outputs": [],
   "source": [
    "forpredtabmnb = pd.DataFrame(mnb3b.predict(x3b_test_pre))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 379,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pale Ales and Lagers                   1939\n",
       "Strong/Dark Ales                        910\n",
       "Other (Sours, Wild, Holiday styles)     167\n",
       "dtype: int64"
      ]
     },
     "execution_count": 379,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forpredtabmnb.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 380,
   "metadata": {},
   "outputs": [],
   "source": [
    "forpredtabfor = pd.DataFrame(forest3b.predict(x3b_test_pre))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 381,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pale Ales and Lagers                   1645\n",
       "Strong/Dark Ales                       1267\n",
       "Other (Sours, Wild, Holiday styles)     104\n",
       "dtype: int64"
      ]
     },
     "execution_count": 381,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forpredtabfor.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 382,
   "metadata": {},
   "outputs": [],
   "source": [
    "forpredtabknn = pd.DataFrame(knn3b.predict(x3b_test_pre))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 383,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pale Ales and Lagers                   1855\n",
       "Strong/Dark Ales                       1127\n",
       "Other (Sours, Wild, Holiday styles)      34\n",
       "dtype: int64"
      ]
     },
     "execution_count": 383,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forpredtabknn.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 384,
   "metadata": {},
   "outputs": [],
   "source": [
    "forpredtrues = pd.DataFrame(y3b_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('TRUES',\n",
       " threesty                           \n",
       " Pale Ales and Lagers                   1609\n",
       " Strong/Dark Ales                       1151\n",
       " Other (Sours, Wild, Holiday styles)     335\n",
       " dtype: int64)"
      ]
     },
     "execution_count": 227,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"TRUES\", forpredtrues.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 550,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pale Ales and Lagers                   13697\n",
       "Strong/Dark Ales                       10577\n",
       "Other (Sours, Wild, Holiday styles)     2722\n",
       "Name: threesty, dtype: int64"
      ]
     },
     "execution_count": 550,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.threesty.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import precision_score, recall_score, precision_recall_curve,f1_score, f1_score, accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Target is multiclass but average='binary'. Please choose another average setting, one of [None, 'micro', 'macro', 'weighted'].",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-217-46cc6f767330>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mylabel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Actual'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtitle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'kNN3b confusion matrix'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 11\u001b[0;31m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Precision Score: \"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mprecision_score\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my3b_test\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mknn3b\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx3b_test_pre\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     12\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Recall Score: \"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrecall_score\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my3b_test\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mknn3b\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx3b_test_pre\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Accuracy : \"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maccuracy_score\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my3b_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mknn3b\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx3b_test_pre\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m;\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/utils/validation.py\u001b[0m in \u001b[0;36minner_f\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m     70\u001b[0m                           FutureWarning)\n\u001b[1;32m     71\u001b[0m         \u001b[0mkwargs\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mupdate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0mk\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0marg\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msig\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mparameters\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 72\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     73\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0minner_f\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     74\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/metrics/_classification.py\u001b[0m in \u001b[0;36mprecision_score\u001b[0;34m(y_true, y_pred, labels, pos_label, average, sample_weight, zero_division)\u001b[0m\n\u001b[1;32m   1615\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1616\u001b[0m     \"\"\"\n\u001b[0;32m-> 1617\u001b[0;31m     p, _, _, _ = precision_recall_fscore_support(y_true, y_pred,\n\u001b[0m\u001b[1;32m   1618\u001b[0m                                                  \u001b[0mlabels\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlabels\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1619\u001b[0m                                                  \u001b[0mpos_label\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mpos_label\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/utils/validation.py\u001b[0m in \u001b[0;36minner_f\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m     70\u001b[0m                           FutureWarning)\n\u001b[1;32m     71\u001b[0m         \u001b[0mkwargs\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mupdate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0mk\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0marg\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msig\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mparameters\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 72\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     73\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0minner_f\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     74\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/metrics/_classification.py\u001b[0m in \u001b[0;36mprecision_recall_fscore_support\u001b[0;34m(y_true, y_pred, beta, labels, pos_label, average, warn_for, sample_weight, zero_division)\u001b[0m\n\u001b[1;32m   1431\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mbeta\u001b[0m \u001b[0;34m<\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1432\u001b[0m         \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"beta should be >=0 in the F-beta score\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1433\u001b[0;31m     labels = _check_set_wise_labels(y_true, y_pred, average, labels,\n\u001b[0m\u001b[1;32m   1434\u001b[0m                                     pos_label)\n\u001b[1;32m   1435\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/metrics/_classification.py\u001b[0m in \u001b[0;36m_check_set_wise_labels\u001b[0;34m(y_true, y_pred, average, labels, pos_label)\u001b[0m\n\u001b[1;32m   1261\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0my_type\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'multiclass'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1262\u001b[0m                 \u001b[0maverage_options\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mremove\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'samples'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1263\u001b[0;31m             raise ValueError(\"Target is %s but average='binary'. Please \"\n\u001b[0m\u001b[1;32m   1264\u001b[0m                              \u001b[0;34m\"choose another average setting, one of %r.\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1265\u001b[0m                              % (y_type, average_options))\n",
      "\u001b[0;31mValueError\u001b[0m: Target is multiclass but average='binary'. Please choose another average setting, one of [None, 'micro', 'macro', 'weighted']."
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Print confusion matrix for kNN\n",
    "knn3b_confusion = confusion_matrix(y3b_test, knn3b.predict(x3b_test_pre))\n",
    "plt.figure(dpi=150)\n",
    "sns.heatmap(knn3b_confusion, cmap=plt.cm.GnBu, annot=True, square=True,\n",
    "           xticklabels=['Other', 'Pale', 'Dark'],\n",
    "           yticklabels=['Other', 'Pale', 'Dark'], fmt = 'g')\n",
    "\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Actual')\n",
    "plt.title('kNN3b confusion matrix')\n",
    "print(\"Precision Score: \", precision_score(y3b_test,knn3b.predict(x3b_test_pre)))\n",
    "print(\"Recall Score: \", recall_score(y3b_test,knn3b.predict(x3b_test_pre)))\n",
    "print(\"Accuracy : \", accuracy_score(y3b_test, knn3b.predict(x3b_test_pre)));"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predicting Quants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 386,
   "metadata": {},
   "outputs": [],
   "source": [
    "corr_df = df[['ipa', 'threesty', 'foursty', 'og', 'fg', 'abv', 'srm', 'ibu', 'bugu', 'hopcmp']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ipa</th>\n",
       "      <th>og</th>\n",
       "      <th>fg</th>\n",
       "      <th>abv</th>\n",
       "      <th>srm</th>\n",
       "      <th>ibu</th>\n",
       "      <th>bugu</th>\n",
       "      <th>hopcmp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ipa</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.037788</td>\n",
       "      <td>0.030705</td>\n",
       "      <td>0.155020</td>\n",
       "      <td>-0.170530</td>\n",
       "      <td>0.514471</td>\n",
       "      <td>0.437217</td>\n",
       "      <td>0.406334</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>og</th>\n",
       "      <td>0.037788</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.980055</td>\n",
       "      <td>0.180430</td>\n",
       "      <td>0.060313</td>\n",
       "      <td>0.052124</td>\n",
       "      <td>-0.012195</td>\n",
       "      <td>0.024153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fg</th>\n",
       "      <td>0.030705</td>\n",
       "      <td>0.980055</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.159587</td>\n",
       "      <td>0.106518</td>\n",
       "      <td>0.057145</td>\n",
       "      <td>-0.015861</td>\n",
       "      <td>0.015112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>abv</th>\n",
       "      <td>0.155020</td>\n",
       "      <td>0.180430</td>\n",
       "      <td>0.159587</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.281197</td>\n",
       "      <td>0.265004</td>\n",
       "      <td>-0.017909</td>\n",
       "      <td>0.112851</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>srm</th>\n",
       "      <td>-0.170530</td>\n",
       "      <td>0.060313</td>\n",
       "      <td>0.106518</td>\n",
       "      <td>0.281197</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.044608</td>\n",
       "      <td>-0.041714</td>\n",
       "      <td>-0.122126</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ibu</th>\n",
       "      <td>0.514471</td>\n",
       "      <td>0.052124</td>\n",
       "      <td>0.057145</td>\n",
       "      <td>0.265004</td>\n",
       "      <td>0.044608</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.859581</td>\n",
       "      <td>0.339367</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bugu</th>\n",
       "      <td>0.437217</td>\n",
       "      <td>-0.012195</td>\n",
       "      <td>-0.015861</td>\n",
       "      <td>-0.017909</td>\n",
       "      <td>-0.041714</td>\n",
       "      <td>0.859581</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.287475</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>hopcmp</th>\n",
       "      <td>0.406334</td>\n",
       "      <td>0.024153</td>\n",
       "      <td>0.015112</td>\n",
       "      <td>0.112851</td>\n",
       "      <td>-0.122126</td>\n",
       "      <td>0.339367</td>\n",
       "      <td>0.287475</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             ipa        og        fg       abv       srm       ibu      bugu  \\\n",
       "ipa     1.000000  0.037788  0.030705  0.155020 -0.170530  0.514471  0.437217   \n",
       "og      0.037788  1.000000  0.980055  0.180430  0.060313  0.052124 -0.012195   \n",
       "fg      0.030705  0.980055  1.000000  0.159587  0.106518  0.057145 -0.015861   \n",
       "abv     0.155020  0.180430  0.159587  1.000000  0.281197  0.265004 -0.017909   \n",
       "srm    -0.170530  0.060313  0.106518  0.281197  1.000000  0.044608 -0.041714   \n",
       "ibu     0.514471  0.052124  0.057145  0.265004  0.044608  1.000000  0.859581   \n",
       "bugu    0.437217 -0.012195 -0.015861 -0.017909 -0.041714  0.859581  1.000000   \n",
       "hopcmp  0.406334  0.024153  0.015112  0.112851 -0.122126  0.339367  0.287475   \n",
       "\n",
       "          hopcmp  \n",
       "ipa     0.406334  \n",
       "og      0.024153  \n",
       "fg      0.015112  \n",
       "abv     0.112851  \n",
       "srm    -0.122126  \n",
       "ibu     0.339367  \n",
       "bugu    0.287475  \n",
       "hopcmp  1.000000  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr_df.corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 387,
   "metadata": {},
   "outputs": [],
   "source": [
    "corr_df = corr_df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 347,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = corr_df.drop(columns =['ipa', 'threesty', 'foursty'])\n",
    "y= corr_df['ipa']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, x_test, y_train, y_test = train_test_split(X, y, test_size=0.2)\n",
    "X_train, x_val, y_train, y_val = train_test_split(X_train, y_train, test_size=0.25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_pre = preprocessing.scale(X)\n",
    "X_train_pre = preprocessing.scale(X_train)\n",
    "x_test_pre = preprocessing.scale(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_depth=100, n_estimators=800)"
      ]
     },
     "execution_count": 164,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forestC = RandomForestClassifier(n_estimators = 800, min_samples_split = 2, min_samples_leaf = 1, max_features = 'auto', max_depth =  100, bootstrap = True)\n",
    "forestC.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 165,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forestC.score(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9088300220750551"
      ]
     },
     "execution_count": 166,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forestC.score(x_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Let's Random Search CV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "# Number of trees in random forest\n",
    "n_estimators = [int(x) for x in np.linspace(start = 200, stop = 2000, num = 10)]\n",
    "# Number of features to consider at every split\n",
    "max_features = ['auto', 'sqrt']\n",
    "# Maximum number of levels in tree\n",
    "max_depth = [int(x) for x in np.linspace(10, 110, num = 11)]\n",
    "max_depth.append(None)\n",
    "# Minimum number of samples required to split a node\n",
    "min_samples_split = [2, 5, 10]\n",
    "# Minimum number of samples required at each leaf node\n",
    "min_samples_leaf = [1, 2, 4]\n",
    "# Method of selecting samples for training each tree\n",
    "bootstrap = [True, False]\n",
    "# Create the random grid\n",
    "random_grid = {'n_estimators': n_estimators,\n",
    "               'max_features': max_features,\n",
    "               'max_depth': max_depth,\n",
    "               'min_samples_split': min_samples_split,\n",
    "               'min_samples_leaf': min_samples_leaf,\n",
    "               'bootstrap': bootstrap}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 100 candidates, totalling 300 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed:  1.9min\n",
      "[Parallel(n_jobs=-1)]: Done 146 tasks      | elapsed: 10.3min\n",
      "[Parallel(n_jobs=-1)]: Done 300 out of 300 | elapsed: 21.6min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=3, estimator=RandomForestClassifier(), n_iter=100,\n",
       "                   n_jobs=-1,\n",
       "                   param_distributions={'bootstrap': [True, False],\n",
       "                                        'max_depth': [10, 20, 30, 40, 50, 60,\n",
       "                                                      70, 80, 90, 100, 110,\n",
       "                                                      None],\n",
       "                                        'max_features': ['auto', 'sqrt'],\n",
       "                                        'min_samples_leaf': [1, 2, 4],\n",
       "                                        'min_samples_split': [2, 5, 10],\n",
       "                                        'n_estimators': [200, 400, 600, 800,\n",
       "                                                         1000, 1200, 1400, 1600,\n",
       "                                                         1800, 2000]},\n",
       "                   random_state=42, verbose=2)"
      ]
     },
     "execution_count": 156,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Use the random grid to search for best hyperparameters\n",
    "# First create the base model to tune\n",
    "# Random search of parameters, using 3 fold cross validation, \n",
    "# search across 100 different combinations, and use all available cores\n",
    "rf_random = RandomizedSearchCV(estimator = forestC, param_distributions = random_grid, n_iter = 100, cv = 3, verbose=2, random_state=42, n_jobs = -1)\n",
    "# Fit the random search model\n",
    "rf_random.fit(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'n_estimators': 800,\n",
       " 'min_samples_split': 2,\n",
       " 'min_samples_leaf': 1,\n",
       " 'max_features': 'auto',\n",
       " 'max_depth': 100,\n",
       " 'bootstrap': True}"
      ]
     },
     "execution_count": 157,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_random.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from get_stats import get_stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "0\t1.047\t1.014\t4.27\t10.07\t20.80\t0.442553\t18.9"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "metadata": {},
   "outputs": [
    {
     "name": "stdin",
     "output_type": "stream",
     "text": [
      "Enter Original Gravity:  1.047\n",
      "Enter Final Gravity:  1.014\n",
      "Enter ABV:  4.27\n",
      "Enter SRM:  10.07\n",
      "Enter IBU:  20.80\n",
      "On a scale of 1-100, how hoppy is this beer? :  18.9\n"
     ]
    }
   ],
   "source": [
    "test_df = get_stats()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "from collections import defaultdict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = [knnq.predict(test_df)[0], svcq.predict(test_df)[0], bnbq.predict(test_df)[0], logq.predict(test_df)[0]] #forestC.predict(test_df)[0]]\n",
    "clean_pred = defaultdict(int)\n",
    "for pred in preds:\n",
    "    if pred == 1:\n",
    "        clean_pred['IPA'] +=1\n",
    "    if pred == 0:\n",
    "        clean_pred['NOT IPA'] += 1\n",
    "res_dic = dict(clean_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({'IPA': 3, 'NOT IPA': 1}, [1, 0, 1, 1])"
      ]
     },
     "execution_count": 244,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res_dic, preds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "IPA\n"
     ]
    }
   ],
   "source": [
    "for key, value in res_dic.items():\n",
    "    if value == len(preds)/2:\n",
    "        print(\"TIE\")\n",
    "        break\n",
    "    elif value > len(preds)/2:\n",
    "        print(key)\n",
    "    else:\n",
    "        continue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 242,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NOT IPA\n"
     ]
    }
   ],
   "source": [
    "svcpred = svcq.predict(test_df)[0]\n",
    "if svcpred == 0:\n",
    "    print('NOT IPA')\n",
    "elif svcpred == 1:\n",
    "    print('IPA')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ipa</th>\n",
       "      <th>og</th>\n",
       "      <th>fg</th>\n",
       "      <th>abv</th>\n",
       "      <th>srm</th>\n",
       "      <th>ibu</th>\n",
       "      <th>bugu</th>\n",
       "      <th>hopcmp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21222</th>\n",
       "      <td>0</td>\n",
       "      <td>1.054</td>\n",
       "      <td>1.013</td>\n",
       "      <td>5.29</td>\n",
       "      <td>16.03</td>\n",
       "      <td>22.07</td>\n",
       "      <td>0.408704</td>\n",
       "      <td>18.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14616</th>\n",
       "      <td>1</td>\n",
       "      <td>1.059</td>\n",
       "      <td>1.011</td>\n",
       "      <td>6.19</td>\n",
       "      <td>6.73</td>\n",
       "      <td>43.00</td>\n",
       "      <td>0.728814</td>\n",
       "      <td>59.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15947</th>\n",
       "      <td>0</td>\n",
       "      <td>1.040</td>\n",
       "      <td>1.009</td>\n",
       "      <td>4.04</td>\n",
       "      <td>48.83</td>\n",
       "      <td>20.17</td>\n",
       "      <td>0.504250</td>\n",
       "      <td>8.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3969</th>\n",
       "      <td>0</td>\n",
       "      <td>1.050</td>\n",
       "      <td>1.013</td>\n",
       "      <td>4.82</td>\n",
       "      <td>3.30</td>\n",
       "      <td>8.18</td>\n",
       "      <td>0.163600</td>\n",
       "      <td>14.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9581</th>\n",
       "      <td>0</td>\n",
       "      <td>1.055</td>\n",
       "      <td>1.013</td>\n",
       "      <td>5.52</td>\n",
       "      <td>4.04</td>\n",
       "      <td>35.92</td>\n",
       "      <td>0.653091</td>\n",
       "      <td>35.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4358</th>\n",
       "      <td>0</td>\n",
       "      <td>1.052</td>\n",
       "      <td>1.013</td>\n",
       "      <td>5.07</td>\n",
       "      <td>13.53</td>\n",
       "      <td>19.06</td>\n",
       "      <td>0.366538</td>\n",
       "      <td>9.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16530</th>\n",
       "      <td>0</td>\n",
       "      <td>1.052</td>\n",
       "      <td>1.010</td>\n",
       "      <td>5.50</td>\n",
       "      <td>9.21</td>\n",
       "      <td>36.87</td>\n",
       "      <td>0.709038</td>\n",
       "      <td>41.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13722</th>\n",
       "      <td>0</td>\n",
       "      <td>1.053</td>\n",
       "      <td>1.008</td>\n",
       "      <td>5.96</td>\n",
       "      <td>4.81</td>\n",
       "      <td>27.66</td>\n",
       "      <td>0.521887</td>\n",
       "      <td>27.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1785</th>\n",
       "      <td>1</td>\n",
       "      <td>1.086</td>\n",
       "      <td>1.020</td>\n",
       "      <td>8.62</td>\n",
       "      <td>8.89</td>\n",
       "      <td>100.24</td>\n",
       "      <td>1.165581</td>\n",
       "      <td>59.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16319</th>\n",
       "      <td>0</td>\n",
       "      <td>1.047</td>\n",
       "      <td>1.014</td>\n",
       "      <td>4.27</td>\n",
       "      <td>10.07</td>\n",
       "      <td>20.80</td>\n",
       "      <td>0.442553</td>\n",
       "      <td>18.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4524</th>\n",
       "      <td>0</td>\n",
       "      <td>1.047</td>\n",
       "      <td>1.009</td>\n",
       "      <td>4.91</td>\n",
       "      <td>5.12</td>\n",
       "      <td>19.04</td>\n",
       "      <td>0.405106</td>\n",
       "      <td>16.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16037</th>\n",
       "      <td>0</td>\n",
       "      <td>1.055</td>\n",
       "      <td>1.019</td>\n",
       "      <td>4.79</td>\n",
       "      <td>40.55</td>\n",
       "      <td>28.92</td>\n",
       "      <td>0.525818</td>\n",
       "      <td>33.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12894</th>\n",
       "      <td>1</td>\n",
       "      <td>1.065</td>\n",
       "      <td>1.012</td>\n",
       "      <td>7.02</td>\n",
       "      <td>22.47</td>\n",
       "      <td>59.24</td>\n",
       "      <td>0.911385</td>\n",
       "      <td>44.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19036</th>\n",
       "      <td>0</td>\n",
       "      <td>1.053</td>\n",
       "      <td>1.009</td>\n",
       "      <td>5.76</td>\n",
       "      <td>6.25</td>\n",
       "      <td>46.04</td>\n",
       "      <td>0.868679</td>\n",
       "      <td>40.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21485</th>\n",
       "      <td>0</td>\n",
       "      <td>1.054</td>\n",
       "      <td>1.009</td>\n",
       "      <td>5.84</td>\n",
       "      <td>5.88</td>\n",
       "      <td>45.61</td>\n",
       "      <td>0.844630</td>\n",
       "      <td>74.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21327</th>\n",
       "      <td>1</td>\n",
       "      <td>1.059</td>\n",
       "      <td>1.011</td>\n",
       "      <td>6.23</td>\n",
       "      <td>4.12</td>\n",
       "      <td>26.09</td>\n",
       "      <td>0.442203</td>\n",
       "      <td>43.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22965</th>\n",
       "      <td>0</td>\n",
       "      <td>1.070</td>\n",
       "      <td>1.019</td>\n",
       "      <td>6.58</td>\n",
       "      <td>4.12</td>\n",
       "      <td>66.41</td>\n",
       "      <td>0.948714</td>\n",
       "      <td>66.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2762</th>\n",
       "      <td>0</td>\n",
       "      <td>1.053</td>\n",
       "      <td>1.015</td>\n",
       "      <td>5.02</td>\n",
       "      <td>10.90</td>\n",
       "      <td>18.91</td>\n",
       "      <td>0.356792</td>\n",
       "      <td>33.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12419</th>\n",
       "      <td>0</td>\n",
       "      <td>1.050</td>\n",
       "      <td>1.009</td>\n",
       "      <td>5.39</td>\n",
       "      <td>7.91</td>\n",
       "      <td>22.42</td>\n",
       "      <td>0.448400</td>\n",
       "      <td>27.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16654</th>\n",
       "      <td>1</td>\n",
       "      <td>1.063</td>\n",
       "      <td>1.015</td>\n",
       "      <td>6.38</td>\n",
       "      <td>12.97</td>\n",
       "      <td>70.85</td>\n",
       "      <td>1.124603</td>\n",
       "      <td>26.1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       ipa     og     fg   abv    srm     ibu      bugu  hopcmp\n",
       "21222    0  1.054  1.013  5.29  16.03   22.07  0.408704    18.9\n",
       "14616    1  1.059  1.011  6.19   6.73   43.00  0.728814    59.5\n",
       "15947    0  1.040  1.009  4.04  48.83   20.17  0.504250     8.7\n",
       "3969     0  1.050  1.013  4.82   3.30    8.18  0.163600    14.5\n",
       "9581     0  1.055  1.013  5.52   4.04   35.92  0.653091    35.5\n",
       "4358     0  1.052  1.013  5.07  13.53   19.06  0.366538     9.1\n",
       "16530    0  1.052  1.010  5.50   9.21   36.87  0.709038    41.1\n",
       "13722    0  1.053  1.008  5.96   4.81   27.66  0.521887    27.7\n",
       "1785     1  1.086  1.020  8.62   8.89  100.24  1.165581    59.5\n",
       "16319    0  1.047  1.014  4.27  10.07   20.80  0.442553    18.9\n",
       "4524     0  1.047  1.009  4.91   5.12   19.04  0.405106    16.2\n",
       "16037    0  1.055  1.019  4.79  40.55   28.92  0.525818    33.2\n",
       "12894    1  1.065  1.012  7.02  22.47   59.24  0.911385    44.8\n",
       "19036    0  1.053  1.009  5.76   6.25   46.04  0.868679    40.8\n",
       "21485    0  1.054  1.009  5.84   5.88   45.61  0.844630    74.9\n",
       "21327    1  1.059  1.011  6.23   4.12   26.09  0.442203    43.1\n",
       "22965    0  1.070  1.019  6.58   4.12   66.41  0.948714    66.4\n",
       "2762     0  1.053  1.015  5.02  10.90   18.91  0.356792    33.4\n",
       "12419    0  1.050  1.009  5.39   7.91   22.42  0.448400    27.3\n",
       "16654    1  1.063  1.015  6.38  12.97   70.85  1.124603    26.1"
      ]
     },
     "execution_count": 221,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr_df.sample(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "11048    1\n",
       "87       0\n",
       "18997    1\n",
       "24449    0\n",
       "23524    0\n",
       "4048     0\n",
       "6042     0\n",
       "14338    0\n",
       "18037    0\n",
       "18400    0\n",
       "17156    0\n",
       "19321    0\n",
       "13341    0\n",
       "22511    0\n",
       "22443    0\n",
       "15302    0\n",
       "14297    0\n",
       "3845     1\n",
       "6182     1\n",
       "22727    0\n",
       "Name: ipa, dtype: int64"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test.tail(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 1, 1, ..., 1, 0, 0])"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forestC.predict(x_test_pre)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "13050     17.6\n",
       "22207     10.0\n",
       "21249     45.8\n",
       "18906     63.4\n",
       "8360      17.5\n",
       "20784     35.5\n",
       "20443    109.7\n",
       "317       54.0\n",
       "10284     26.1\n",
       "5689       8.8\n",
       "15771      9.0\n",
       "22543     37.4\n",
       "9364       9.0\n",
       "2320      17.4\n",
       "17325     17.8\n",
       "10899      9.1\n",
       "16174      8.8\n",
       "4781      25.4\n",
       "1843      57.2\n",
       "115       18.1\n",
       "Name: hopcmp, dtype: float64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr_df.hopcmp.sample(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "395.5"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr_df.hopcmp.max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "norms = []\n",
    "for index, row in corr_df.iterrows():\n",
    "    norm = (row.hopcmp - 1)/ 200\n",
    "    norms.append(norm)\n",
    "corr_df['normcmp'] = norms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0        0.4635\n",
       "1        0.1240\n",
       "2        0.1945\n",
       "3        0.2440\n",
       "4        0.4355\n",
       "          ...  \n",
       "24967    0.1800\n",
       "24968    0.1915\n",
       "24969    0.1915\n",
       "24970    0.0935\n",
       "24971    0.0935\n",
       "Name: normcmp, Length: 22648, dtype: float64"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr_df.normcmp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(n_neighbors=20)"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knnq = KNeighborsClassifier(n_neighbors = 20)\n",
    "knnq.fit(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9025610833088019"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knnq.score(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9013245033112582"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knnq.score(x_test_pre,y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.9009420076538122, 0.9028697571743929)"
      ]
     },
     "execution_count": 146,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svcq = SVC(probability = True)\n",
    "svcq.fit(X_train_pre, y_train)\n",
    "svcq.score(X_train_pre, y_train), svcq.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8713024282560706"
      ]
     },
     "execution_count": 144,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logq = LogisticRegression()\n",
    "logq.fit(X_train_pre, y_train)\n",
    "logq.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "import xgboost as xgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8969094922737307"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbq = xgb.XGBClassifier(learning_rate = .3, n_estimators = 100 ,max_depth = 4, min_child_weight = 1)\n",
    "xgbq.fit(X_train_pre, y_train)\n",
    "xgbq.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.7962908448631145, 0.7902869757174393)"
      ]
     },
     "execution_count": 152,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bnbq = MultinomialNB()\n",
    "bnbq.fit(abs(X_train_pre), y_train)\n",
    "bnbq.score(abs(X_train_pre), y_train), bnbq.score(abs(x_test_pre), y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5761910312655648"
      ]
     },
     "execution_count": 140,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[df.ipa == 0].bugu.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve, roc_auc_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [],
   "source": [
    "fpr, tpr, thresholds = roc_curve(y_test, knnq.predict_proba(x_test_pre)[:,1])\n",
    "fprlog, tprlog, thresholdslog = roc_curve(y_test, logq.predict_proba(x_test_pre)[:,1])\n",
    "fprnb, tprnb, thresholdsnb = roc_curve(y_test, bnbq.predict_proba(x_test_pre)[:,1])\n",
    "fprrf, tprrf, thresholdsrf = roc_curve(y_test, forestC.predict_proba(x_test_pre)[:,1])\n",
    "\n",
    "fprsvc, tprsvc, thresholdssvc = roc_curve(y_test, svcq.predict_proba(x_test_pre)[:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KNN ROC AUC score =  0.9347476142947563\n",
      "Logit ROC AUC score =  0.9196965639647853\n",
      "Naive Bayes ROC AUC score =  0.766370740365336\n",
      "Random Forest ROC AUC score =  0.48635647150210903\n",
      "SVC ROC AUC score =  0.9198296256788637\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (15,10))\n",
    "plt.plot(fpr, tpr,lw=2)\n",
    "plt.plot(fprlog, tprlog,lw=2)\n",
    "plt.plot(fprnb, tprnb,lw=2)\n",
    "plt.plot(fprrf, tprrf,lw=2)\n",
    "\n",
    "plt.plot(fprsvc, tprsvc,lw=2)\n",
    "plt.plot([0,1],[0,1],c='violet',ls='--')\n",
    "plt.xlim([-0.05,1.05])\n",
    "plt.ylim([-0.05,1.05])\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC Scores', fontweight=\"bold\", fontsize=20);\n",
    "plt.legend(['KNN','Logit', 'Bernoulli NB', 'Random Forest','SVC'])\n",
    "print(\"KNN ROC AUC score = \", roc_auc_score(y_test, knnq.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"Logit ROC AUC score = \", roc_auc_score(y_test, logq.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"Naive Bayes ROC AUC score = \", roc_auc_score(y_test, bnbq.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"Random Forest ROC AUC score = \", roc_auc_score(y_test, forestC.predict_proba(x_test_pre)[:,1]))\n",
    "print(\"SVC ROC AUC score = \", roc_auc_score(y_test, svcq.predict_proba(x_test_pre)[:,1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['ipa', 'threesty', 'og', 'fg', 'abv', 'srm', 'ibu', 'bugu', 'hopcmp'], dtype='object')"
      ]
     },
     "execution_count": 248,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr_df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 352,
   "metadata": {},
   "outputs": [],
   "source": [
    "corr_df = corr_df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 388,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_ = corr_df.drop(columns = ['ipa', 'threesty', 'foursty'])\n",
    "y_ = corr_df['foursty']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 389,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, x_test, y_train, y_test = train_test_split(X_, y_, test_size=0.2)\n",
    "X_train, x_val, y_train, y_val = train_test_split(X_train, y_train, test_size=0.25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 390,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_pre = preprocessing.scale(X)\n",
    "X_train_pre = preprocessing.scale(X_train)\n",
    "x_test_pre = preprocessing.scale(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 391,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_depth=20, max_features='sqrt', min_samples_leaf=2,\n",
       "                       n_estimators=461)"
      ]
     },
     "execution_count": 391,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest3 = RandomForestClassifier(n_estimators = 461, min_samples_split= 2, min_samples_leaf=2, max_features='sqrt',\\\n",
    "                                  max_depth = 20, bootstrap = True)\n",
    "forest3.fit(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 392,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9374448042390344"
      ]
     },
     "execution_count": 392,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest3.score(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 393,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6781456953642384"
      ]
     },
     "execution_count": 393,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest3.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "# Number of trees in random forest\n",
    "n_estimators = [int(x) for x in np.linspace(start = 30, stop = 1000, num = 10)]\n",
    "# Number of features to consider at every split\n",
    "max_features = ['auto', 'sqrt']\n",
    "# Maximum number of levels in tree\n",
    "max_depth = [int(x) for x in np.linspace(10, 110, num = 11)]\n",
    "max_depth.append(None)\n",
    "# Minimum number of samples required to split a node\n",
    "min_samples_split = [2, 5, 10]\n",
    "# Minimum number of samples required at each leaf node\n",
    "min_samples_leaf = [1, 2, 4]\n",
    "# Method of selecting samples for training each tree\n",
    "bootstrap = [True, False]\n",
    "# Create the random grid\n",
    "random_grid = {'n_estimators': n_estimators,\n",
    "               'max_features': max_features,\n",
    "               'max_depth': max_depth,\n",
    "               'min_samples_split': min_samples_split,\n",
    "               'min_samples_leaf': min_samples_leaf,\n",
    "               'bootstrap': bootstrap}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 268,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 100 candidates, totalling 300 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed:   54.1s\n",
      "[Parallel(n_jobs=-1)]: Done 146 tasks      | elapsed:  4.9min\n",
      "[Parallel(n_jobs=-1)]: Done 300 out of 300 | elapsed:  9.8min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=3,\n",
       "                   estimator=RandomForestClassifier(max_depth=100,\n",
       "                                                    n_estimators=800),\n",
       "                   n_iter=100, n_jobs=-1,\n",
       "                   param_distributions={'bootstrap': [True, False],\n",
       "                                        'max_depth': [10, 20, 30, 40, 50, 60,\n",
       "                                                      70, 80, 90, 100, 110,\n",
       "                                                      None],\n",
       "                                        'max_features': ['auto', 'sqrt'],\n",
       "                                        'min_samples_leaf': [1, 2, 4],\n",
       "                                        'min_samples_split': [2, 5, 10],\n",
       "                                        'n_estimators': [30, 137, 245, 353, 461,\n",
       "                                                         568, 676, 784, 892,\n",
       "                                                         1000]},\n",
       "                   random_state=42, verbose=2)"
      ]
     },
     "execution_count": 268,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Use the random grid to search for best hyperparameters\n",
    "# First create the base model to tune\n",
    "# Random search of parameters, using 3 fold cross validation, \n",
    "# search across 100 different combinations, and use all available cores\n",
    "rf_random = RandomizedSearchCV(estimator = forest3, param_distributions = random_grid, n_iter = 100, cv = 3, verbose=2, random_state=42, n_jobs = -1)\n",
    "# Fit the random search model\n",
    "rf_random.fit(X_train_pre, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 274,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'n_estimators': 461,\n",
       " 'min_samples_split': 2,\n",
       " 'min_samples_leaf': 1,\n",
       " 'max_features': 'sqrt',\n",
       " 'max_depth': 20,\n",
       " 'bootstrap': True}"
      ]
     },
     "execution_count": 274,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_random.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 394,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_depth=20, max_features='sqrt', min_samples_leaf=2,\n",
       "                       n_estimators=461)"
      ]
     },
     "execution_count": 394,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest3op = RandomForestClassifier(n_estimators = 461, min_samples_split= 2, min_samples_leaf=2, max_features='sqrt',\\\n",
    "                                  max_depth = 20, bootstrap = True)\n",
    "forest3op.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 395,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7185430463576159"
      ]
     },
     "execution_count": 395,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest3op.score(x_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 396,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6821192052980133"
      ]
     },
     "execution_count": 396,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn3 = KNeighborsClassifier(n_neighbors = 90)\n",
    "knn3.fit(X_train_pre, y_train)\n",
    "knn3.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 397,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5591611479028697"
      ]
     },
     "execution_count": 397,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnb3 = MultinomialNB()\n",
    "mnb3.fit(X_train, y_train)\n",
    "mnb3.score(x_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 398,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6214128035320088"
      ]
     },
     "execution_count": 398,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logr3 = LogisticRegression()\n",
    "logr3.fit(X_train_pre, y_train)\n",
    "logr3.score(x_test_pre, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 561,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7092715231788079"
      ]
     },
     "execution_count": 561,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb3 = xgb.XGBClassifier()\n",
    "xgb3.fit(X_train, y_train)\n",
    "xgb3.score(x_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 521,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Strong/Dark Ales'], dtype=object)"
      ]
     },
     "execution_count": 521,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb3.predict(test_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 368,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NOT IPA\n"
     ]
    }
   ],
   "source": [
    "svcpred = svcq.predict(test_df)[0]\n",
    "if svcpred == 0:\n",
    "    print('NOT IPA')\n",
    "elif svcpred == 1:\n",
    "    print('IPA')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 517,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 564,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ipa</th>\n",
       "      <th>threesty</th>\n",
       "      <th>foursty</th>\n",
       "      <th>og</th>\n",
       "      <th>fg</th>\n",
       "      <th>abv</th>\n",
       "      <th>srm</th>\n",
       "      <th>ibu</th>\n",
       "      <th>bugu</th>\n",
       "      <th>hopcmp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5045</th>\n",
       "      <td>1</td>\n",
       "      <td>Pale Ales and Lagers</td>\n",
       "      <td>IPA/Pale Ales</td>\n",
       "      <td>1.074</td>\n",
       "      <td>1.021</td>\n",
       "      <td>6.99</td>\n",
       "      <td>7.45</td>\n",
       "      <td>51.09</td>\n",
       "      <td>0.690405</td>\n",
       "      <td>27.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20304</th>\n",
       "      <td>1</td>\n",
       "      <td>Pale Ales and Lagers</td>\n",
       "      <td>IPA/Pale Ales</td>\n",
       "      <td>1.054</td>\n",
       "      <td>1.010</td>\n",
       "      <td>5.77</td>\n",
       "      <td>4.77</td>\n",
       "      <td>88.12</td>\n",
       "      <td>1.631852</td>\n",
       "      <td>90.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11380</th>\n",
       "      <td>0</td>\n",
       "      <td>Pale Ales and Lagers</td>\n",
       "      <td>Lager/Cream Ale</td>\n",
       "      <td>1.056</td>\n",
       "      <td>1.017</td>\n",
       "      <td>5.17</td>\n",
       "      <td>9.52</td>\n",
       "      <td>19.85</td>\n",
       "      <td>0.354464</td>\n",
       "      <td>19.9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>952</th>\n",
       "      <td>0</td>\n",
       "      <td>Pale Ales and Lagers</td>\n",
       "      <td>Lager/Cream Ale</td>\n",
       "      <td>1.055</td>\n",
       "      <td>1.010</td>\n",
       "      <td>5.83</td>\n",
       "      <td>6.84</td>\n",
       "      <td>21.65</td>\n",
       "      <td>0.393636</td>\n",
       "      <td>14.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21882</th>\n",
       "      <td>0</td>\n",
       "      <td>Strong/Dark Ales</td>\n",
       "      <td>Strong/Dark Ales</td>\n",
       "      <td>1.056</td>\n",
       "      <td>1.010</td>\n",
       "      <td>6.09</td>\n",
       "      <td>15.08</td>\n",
       "      <td>21.63</td>\n",
       "      <td>0.386250</td>\n",
       "      <td>19.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7887</th>\n",
       "      <td>0</td>\n",
       "      <td>Pale Ales and Lagers</td>\n",
       "      <td>IPA/Pale Ales</td>\n",
       "      <td>1.063</td>\n",
       "      <td>1.014</td>\n",
       "      <td>6.48</td>\n",
       "      <td>11.83</td>\n",
       "      <td>33.77</td>\n",
       "      <td>0.536032</td>\n",
       "      <td>26.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8921</th>\n",
       "      <td>0</td>\n",
       "      <td>Strong/Dark Ales</td>\n",
       "      <td>Strong/Dark Ales</td>\n",
       "      <td>1.065</td>\n",
       "      <td>1.019</td>\n",
       "      <td>6.01</td>\n",
       "      <td>18.38</td>\n",
       "      <td>24.03</td>\n",
       "      <td>0.369692</td>\n",
       "      <td>29.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14239</th>\n",
       "      <td>1</td>\n",
       "      <td>Pale Ales and Lagers</td>\n",
       "      <td>IPA/Pale Ales</td>\n",
       "      <td>1.039</td>\n",
       "      <td>1.011</td>\n",
       "      <td>3.70</td>\n",
       "      <td>8.42</td>\n",
       "      <td>32.84</td>\n",
       "      <td>0.842051</td>\n",
       "      <td>18.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22653</th>\n",
       "      <td>0</td>\n",
       "      <td>Other (Sours, Wild, Holiday styles)</td>\n",
       "      <td>Other (Sours, Wild, Holiday styles)</td>\n",
       "      <td>1.056</td>\n",
       "      <td>1.014</td>\n",
       "      <td>5.47</td>\n",
       "      <td>4.10</td>\n",
       "      <td>19.20</td>\n",
       "      <td>0.342857</td>\n",
       "      <td>7.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18315</th>\n",
       "      <td>0</td>\n",
       "      <td>Strong/Dark Ales</td>\n",
       "      <td>Strong/Dark Ales</td>\n",
       "      <td>1.089</td>\n",
       "      <td>1.025</td>\n",
       "      <td>8.51</td>\n",
       "      <td>40.00</td>\n",
       "      <td>40.04</td>\n",
       "      <td>0.449888</td>\n",
       "      <td>28.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       ipa                             threesty  \\\n",
       "5045     1                 Pale Ales and Lagers   \n",
       "20304    1                 Pale Ales and Lagers   \n",
       "11380    0                 Pale Ales and Lagers   \n",
       "952      0                 Pale Ales and Lagers   \n",
       "21882    0                     Strong/Dark Ales   \n",
       "7887     0                 Pale Ales and Lagers   \n",
       "8921     0                     Strong/Dark Ales   \n",
       "14239    1                 Pale Ales and Lagers   \n",
       "22653    0  Other (Sours, Wild, Holiday styles)   \n",
       "18315    0                     Strong/Dark Ales   \n",
       "\n",
       "                                   foursty     og     fg   abv    srm    ibu  \\\n",
       "5045                         IPA/Pale Ales  1.074  1.021  6.99   7.45  51.09   \n",
       "20304                        IPA/Pale Ales  1.054  1.010  5.77   4.77  88.12   \n",
       "11380                      Lager/Cream Ale  1.056  1.017  5.17   9.52  19.85   \n",
       "952                        Lager/Cream Ale  1.055  1.010  5.83   6.84  21.65   \n",
       "21882                     Strong/Dark Ales  1.056  1.010  6.09  15.08  21.63   \n",
       "7887                         IPA/Pale Ales  1.063  1.014  6.48  11.83  33.77   \n",
       "8921                      Strong/Dark Ales  1.065  1.019  6.01  18.38  24.03   \n",
       "14239                        IPA/Pale Ales  1.039  1.011  3.70   8.42  32.84   \n",
       "22653  Other (Sours, Wild, Holiday styles)  1.056  1.014  5.47   4.10  19.20   \n",
       "18315                     Strong/Dark Ales  1.089  1.025  8.51  40.00  40.04   \n",
       "\n",
       "           bugu  hopcmp  \n",
       "5045   0.690405    27.2  \n",
       "20304  1.631852    90.0  \n",
       "11380  0.354464    19.9  \n",
       "952    0.393636    14.4  \n",
       "21882  0.386250    19.0  \n",
       "7887   0.536032    26.8  \n",
       "8921   0.369692    29.8  \n",
       "14239  0.842051    18.3  \n",
       "22653  0.342857     7.6  \n",
       "18315  0.449888    28.2  "
      ]
     },
     "execution_count": 564,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr_df.sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 520,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>og</th>\n",
       "      <th>fg</th>\n",
       "      <th>abv</th>\n",
       "      <th>srm</th>\n",
       "      <th>ibu</th>\n",
       "      <th>bugu</th>\n",
       "      <th>hopcmp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.02</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     og   fg  abv  srm  ibu  bugu  hopcmp\n",
       "0  1.02  1.0  5.0  3.0  2.0   0.1     2.0"
      ]
     },
     "execution_count": 520,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 519,
   "metadata": {},
   "outputs": [
    {
     "name": "stdin",
     "output_type": "stream",
     "text": [
      "Enter Original Gravity:  1.02\n",
      "Enter Final Gravity:  1\n",
      "Enter ABV:  5\n",
      "Enter SRM:  3\n",
      "Enter IBU:  2\n",
      "On a scale of 1-100, how hoppy is this beer? :  2\n"
     ]
    }
   ],
   "source": [
    "test_df = get_stats()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 556,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = [knn3.predict(test_df)[0], mnb3.predict(test_df)[0], logr3.predict(test_df)[0], forest3op.predict(test_df)[0], forest3.predict(test_df)[0], xgb3.predict(test_df)[0], xgb3.predict(test_df)[0]]\n",
    "clean_pred = defaultdict(int)\n",
    "for pred in preds:\n",
    "    if pred == 'IPA/Pale Ales':\n",
    "        clean_pred['IPA/Pale Ales'] +=1\n",
    "    elif pred == 'Strong/Dark Ales':\n",
    "        clean_pred['Strong/Dark Ales'] += 1\n",
    "    elif pred == 'Lager/Cream Ale':\n",
    "        clean_pred['Lager/Cream Ale'] += 1\n",
    "    else:\n",
    "        clean_pred['Other'] += 1\n",
    "res_dic = dict(clean_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 557,
   "metadata": {
    "jupyter": {
     "source_hidden": true
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({'Strong/Dark Ales': 7},\n",
       " ['Strong/Dark Ales',\n",
       "  'Strong/Dark Ales',\n",
       "  'Strong/Dark Ales',\n",
       "  'Strong/Dark Ales',\n",
       "  'Strong/Dark Ales',\n",
       "  'Strong/Dark Ales',\n",
       "  'Strong/Dark Ales'])"
      ]
     },
     "execution_count": 557,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res_dic, preds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 505,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Strong/Dark Ales\n"
     ]
    }
   ],
   "source": [
    "for key, value in res_dic.items():\n",
    "    if value == len(preds)/2:\n",
    "        print(\"TIE\")\n",
    "        break\n",
    "    elif value > len(preds)//2:\n",
    "        print(key)\n",
    "    else:\n",
    "        continue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 311,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array(['Pale Ales and Lagers'], dtype=object),\n",
       " array(['Pale Ales and Lagers'], dtype=object))"
      ]
     },
     "execution_count": 311,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest3op.predict(test_df), knn3.predict(test_df), "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 492,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 506,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('knn3.pickle', 'wb') as to_write:\n",
    "    pickle.dump(knn3, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 507,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('mnb3.pickle', 'wb') as to_write:\n",
    "    pickle.dump(mnb3, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 508,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('logr3.pickle', 'wb') as to_write:\n",
    "    pickle.dump(logr3, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 510,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('forest3.pickle', 'wb') as to_write:\n",
    "    pickle.dump(forest3, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 511,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('forest3op.pickle', 'wb') as to_write:\n",
    "    pickle.dump(forest3op, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 562,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('web_app/xgb3.pickle', 'wb') as to_write:\n",
    "    pickle.dump(xgb3, to_write)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 525,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best learning_rate: 0.49\n"
     ]
    }
   ],
   "source": [
    "learning_rate = [round(x,2) for x in np.linspace(start = .01, stop = .6, num = 60)]\n",
    "n_estimators = [int(x) for x in np.linspace(start = 100, stop = 2000, num = 20)]\n",
    "max_depth = range(3,10,1)\n",
    "child_weight = range(1,6,2)\n",
    "gamma = [0,.1,.2,.3,.4,.5,.6,.7,.8,.9,1,1.1,1.2,1.3,1.4,1.5,2]\n",
    "subsample = [.6, .7, .8, .9, 1]\n",
    "col_sample = [.6, .7, .8, .9, 1]\n",
    "# Tuning the learning_rate:\n",
    "xgb_tune = xgb.XGBClassifier(n_estimators = 100, max_depth = 3, min_child_weight = 1, subsample = .8, colsample_bytree = 1,gamma = 1, n_jobs = -1)\n",
    "xgb_grid = GridSearchCV(estimator=xgb_tune, param_grid = {'learning_rate':learning_rate},cv=10, verbose = 0, n_jobs = -1)\n",
    "xgb_grid.fit(X_train,y_train)\n",
    "best_learning_rate = xgb_grid.best_params_['learning_rate']\n",
    "print(\"Best learning_rate:\", best_learning_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 527,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best n_estimators: 100\n"
     ]
    }
   ],
   "source": [
    "# Tuning the learning_rate:\n",
    "xgb_tune = xgb.XGBClassifier(learning_rate = best_learning_rate, n_estimators = 100,max_depth = 3, min_child_weight = 1, subsample = .8, colsample_bytree = 1,gamma = 1, n_jobs = -1)\n",
    "xgb_grid = GridSearchCV(estimator=xgb_tune, param_grid = {'n_estimators':n_estimators},cv=10, verbose = 0, n_jobs = -1)\n",
    "xgb_grid.fit(X_train,y_train)\n",
    "best_n_estimators = xgb_grid.best_params_['n_estimators']\n",
    "print(\"Best n_estimators:\", best_n_estimators)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 529,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best max_depth: 5\n"
     ]
    }
   ],
   "source": [
    "xgb_tune = xgb.XGBClassifier(learning_rate = best_learning_rate, n_estimators = best_n_estimators ,max_depth = 3, min_child_weight = 1, subsample = .8, colsample_bytree = 1,gamma = 1, n_jobs = -1)\n",
    "xgb_grid = GridSearchCV(estimator=xgb_tune, param_grid = {'max_depth': max_depth},cv=10, verbose = 0, n_jobs = -1)\n",
    "xgb_grid.fit(X_train,y_train)\n",
    "best_max_depth = xgb_grid.best_params_['max_depth']\n",
    "print(\"Best max_depth:\", best_max_depth)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 532,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best min_child_weight: 1\n"
     ]
    }
   ],
   "source": [
    "xgb_tune = xgb.XGBClassifier(learning_rate = best_learning_rate, n_estimators = best_n_estimators ,max_depth = best_max_depth, min_child_weight = 1, subsample = .8, colsample_bytree = 1,gamma = 1, n_jobs = -1)\n",
    "xgb_grid = GridSearchCV(estimator=xgb_tune, param_grid = {'min_child_weight': child_weight},cv=10, verbose = 0, n_jobs = -1)\n",
    "xgb_grid.fit(X_train,y_train)\n",
    "best_min_child_weight = xgb_grid.best_params_['min_child_weight']\n",
    "print(\"Best min_child_weight:\", best_min_child_weight)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 533,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[12:13:53] WARNING: /Users/runner/miniforge3/conda-bld/xgboost_1598185652448/work/src/learner.cc:516: \n",
      "Parameters: { best_subsample } might not be used.\n",
      "\n",
      "  This may not be accurate due to some parameters are only used in language bindings but\n",
      "  passed down to XGBoost core.  Or some parameters are not used but slip through this\n",
      "  verification. Please open an issue if you find above cases.\n",
      "\n",
      "\n",
      "Best min_child_weight: 1\n"
     ]
    }
   ],
   "source": [
    "xgb_tune = xgb.XGBClassifier(learning_rate = best_learning_rate, n_estimators = best_n_estimators ,max_depth = best_max_depth, min_child_weight = best_min_child_weight, subsample = .8, colsample_bytree = 1,gamma = 1, n_jobs = -1)\n",
    "xgb_grid = GridSearchCV(estimator=xgb_tune, param_grid = {'best_subsample': subsample},cv=10, verbose = 0, n_jobs = -1)\n",
    "xgb_grid.fit(X_train,y_train)\n",
    "best_subsample = xgb_grid.best_params_['best_subsample']\n",
    "print(\"Best min_child_weight:\", best_min_child_weight)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 548,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7075055187637969"
      ]
     },
     "execution_count": 548,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb3x = xgb.XGBClassifier(learning_rate = 0.19, n_estimators = 90, max_depth = 5, subsample = 1)\n",
    "xgb3x.fit(X_train, y_train)\n",
    "xgb3x.score(x_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 568,
   "metadata": {},
   "outputs": [],
   "source": [
    "no_null_notes = df.dropna(subset=['notes'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 569,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true,
     "source_hidden": true
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>og</th>\n",
       "      <th>fg</th>\n",
       "      <th>abv</th>\n",
       "      <th>ibu</th>\n",
       "      <th>srm</th>\n",
       "      <th>Mash pH</th>\n",
       "      <th>IBU (rager)</th>\n",
       "      <th>SRM (daniels)</th>\n",
       "      <th>SRM (ebcmorey)</th>\n",
       "      <th>SRM (mosher)</th>\n",
       "      <th>...</th>\n",
       "      <th>lager</th>\n",
       "      <th>stout</th>\n",
       "      <th>kolsch</th>\n",
       "      <th>wit</th>\n",
       "      <th>puree</th>\n",
       "      <th>lovehop</th>\n",
       "      <th>strong</th>\n",
       "      <th>dry</th>\n",
       "      <th>sweet</th>\n",
       "      <th>bugu</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>16767.000000</td>\n",
       "      <td>16767.000000</td>\n",
       "      <td>16330.000000</td>\n",
       "      <td>16767.000000</td>\n",
       "      <td>15408.000000</td>\n",
       "      <td>1735.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>540.000000</td>\n",
       "      <td>635.000000</td>\n",
       "      <td>184.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>16767.000000</td>\n",
       "      <td>16767.000000</td>\n",
       "      <td>16767.000000</td>\n",
       "      <td>16767.000000</td>\n",
       "      <td>16767.000000</td>\n",
       "      <td>16767.000000</td>\n",
       "      <td>16767.000000</td>\n",
       "      <td>16767.000000</td>\n",
       "      <td>16767.000000</td>\n",
       "      <td>16767.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>1.062235</td>\n",
       "      <td>1.014446</td>\n",
       "      <td>6.255043</td>\n",
       "      <td>44.781897</td>\n",
       "      <td>13.710023</td>\n",
       "      <td>3.710144</td>\n",
       "      <td>NaN</td>\n",
       "      <td>17.472778</td>\n",
       "      <td>27.077543</td>\n",
       "      <td>13.897554</td>\n",
       "      <td>...</td>\n",
       "      <td>0.030238</td>\n",
       "      <td>0.088805</td>\n",
       "      <td>0.012704</td>\n",
       "      <td>0.017892</td>\n",
       "      <td>0.015924</td>\n",
       "      <td>0.000775</td>\n",
       "      <td>0.451721</td>\n",
       "      <td>0.008350</td>\n",
       "      <td>0.705314</td>\n",
       "      <td>0.723803</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.017652</td>\n",
       "      <td>0.005718</td>\n",
       "      <td>1.830069</td>\n",
       "      <td>37.367555</td>\n",
       "      <td>12.157208</td>\n",
       "      <td>2.593348</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10.649507</td>\n",
       "      <td>23.928009</td>\n",
       "      <td>11.835156</td>\n",
       "      <td>...</td>\n",
       "      <td>0.171247</td>\n",
       "      <td>0.284471</td>\n",
       "      <td>0.111995</td>\n",
       "      <td>0.132564</td>\n",
       "      <td>0.125186</td>\n",
       "      <td>0.027835</td>\n",
       "      <td>0.497678</td>\n",
       "      <td>0.090997</td>\n",
       "      <td>0.455915</td>\n",
       "      <td>0.601657</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.991000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.800000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>5.300000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>1.051000</td>\n",
       "      <td>1.011000</td>\n",
       "      <td>5.120000</td>\n",
       "      <td>23.370000</td>\n",
       "      <td>5.190000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9.740000</td>\n",
       "      <td>10.640000</td>\n",
       "      <td>6.130000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.402372</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.059000</td>\n",
       "      <td>1.014000</td>\n",
       "      <td>5.860000</td>\n",
       "      <td>35.210000</td>\n",
       "      <td>8.610000</td>\n",
       "      <td>5.400000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.230000</td>\n",
       "      <td>17.170000</td>\n",
       "      <td>6.965000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.609167</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.069000</td>\n",
       "      <td>1.017000</td>\n",
       "      <td>6.980000</td>\n",
       "      <td>56.440000</td>\n",
       "      <td>17.460000</td>\n",
       "      <td>5.640000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>23.795000</td>\n",
       "      <td>33.115000</td>\n",
       "      <td>15.067500</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.902577</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.260000</td>\n",
       "      <td>1.208000</td>\n",
       "      <td>22.330000</td>\n",
       "      <td>702.290000</td>\n",
       "      <td>87.370000</td>\n",
       "      <td>7.520000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>40.000000</td>\n",
       "      <td>98.500000</td>\n",
       "      <td>40.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>17.230666</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 104 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 og            fg           abv           ibu           srm  \\\n",
       "count  16767.000000  16767.000000  16330.000000  16767.000000  15408.000000   \n",
       "mean       1.062235      1.014446      6.255043     44.781897     13.710023   \n",
       "std        0.017652      0.005718      1.830069     37.367555     12.157208   \n",
       "min        1.000000      0.991000      0.000000      0.000000      0.000000   \n",
       "25%        1.051000      1.011000      5.120000     23.370000      5.190000   \n",
       "50%        1.059000      1.014000      5.860000     35.210000      8.610000   \n",
       "75%        1.069000      1.017000      6.980000     56.440000     17.460000   \n",
       "max        1.260000      1.208000     22.330000    702.290000     87.370000   \n",
       "\n",
       "           Mash pH  IBU (rager)  SRM (daniels)  SRM (ebcmorey)  SRM (mosher)  \\\n",
       "count  1735.000000          0.0     540.000000      635.000000    184.000000   \n",
       "mean      3.710144          NaN      17.472778       27.077543     13.897554   \n",
       "std       2.593348          NaN      10.649507       23.928009     11.835156   \n",
       "min       0.000000          NaN       8.800000        0.000000      5.300000   \n",
       "25%       0.000000          NaN       9.740000       10.640000      6.130000   \n",
       "50%       5.400000          NaN      11.230000       17.170000      6.965000   \n",
       "75%       5.640000          NaN      23.795000       33.115000     15.067500   \n",
       "max       7.520000          NaN      40.000000       98.500000     40.000000   \n",
       "\n",
       "       ...         lager         stout        kolsch           wit  \\\n",
       "count  ...  16767.000000  16767.000000  16767.000000  16767.000000   \n",
       "mean   ...      0.030238      0.088805      0.012704      0.017892   \n",
       "std    ...      0.171247      0.284471      0.111995      0.132564   \n",
       "min    ...      0.000000      0.000000      0.000000      0.000000   \n",
       "25%    ...      0.000000      0.000000      0.000000      0.000000   \n",
       "50%    ...      0.000000      0.000000      0.000000      0.000000   \n",
       "75%    ...      0.000000      0.000000      0.000000      0.000000   \n",
       "max    ...      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "              puree       lovehop        strong           dry         sweet  \\\n",
       "count  16767.000000  16767.000000  16767.000000  16767.000000  16767.000000   \n",
       "mean       0.015924      0.000775      0.451721      0.008350      0.705314   \n",
       "std        0.125186      0.027835      0.497678      0.090997      0.455915   \n",
       "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "25%        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
       "50%        0.000000      0.000000      0.000000      0.000000      1.000000   \n",
       "75%        0.000000      0.000000      1.000000      0.000000      1.000000   \n",
       "max        1.000000      1.000000      1.000000      1.000000      1.000000   \n",
       "\n",
       "               bugu  \n",
       "count  16767.000000  \n",
       "mean       0.723803  \n",
       "std        0.601657  \n",
       "min        0.000000  \n",
       "25%        0.402372  \n",
       "50%        0.609167  \n",
       "75%        0.902577  \n",
       "max       17.230666  \n",
       "\n",
       "[8 rows x 104 columns]"
      ]
     },
     "execution_count": 569,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "no_null_notes.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 672,
   "metadata": {},
   "outputs": [],
   "source": [
    "brq_df = no_null_notes[['foursty', 'dryh', 'lact', 'clone', 'juicy', 'dank', 'wedding', 'whisk', 'trop',\\\n",
    "              'award', 'best', 'smoke', 'roast', 'thick',\\\n",
    "              'heavy', 'stfruit', 'hazy', 'pheno', 'banana','clove', \\\n",
    "              'bitter','puree','lovehop', \\\n",
    "              'strong', 'dry', 'sweet', 'sour', 'dark', 'pine', 'sticky',\\\n",
    "                        'foam', 'hoppy','bisc','ester', 'straw', 'choco', 'coffee', \\\n",
    "                       'og', 'fg', 'abv', 'srm', 'ibu', 'bugu', 'hopcmp']].dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 673,
   "metadata": {},
   "outputs": [],
   "source": [
    "Xbq = brq_df.drop(columns=['foursty'])\n",
    "ybq = brq_df['foursty']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 674,
   "metadata": {},
   "outputs": [],
   "source": [
    "Xb_train, xb_test, yb_train, yb_test = train_test_split(Xbq, ybq, test_size=0.2)\n",
    "Xb_train, xb_val, yb_train, yb_val = train_test_split(Xb_train, yb_train, test_size=0.25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 675,
   "metadata": {},
   "outputs": [],
   "source": [
    "Xb_pre = preprocessing.scale(Xbq)\n",
    "Xb_train_pre = preprocessing.scale(Xb_train)\n",
    "xb_test_pre = preprocessing.scale(xb_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 676,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_depth=20, max_features='sqrt', min_samples_leaf=2,\n",
       "                       n_estimators=150)"
      ]
     },
     "execution_count": 676,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forestbq = RandomForestClassifier(n_estimators = 150, min_samples_split= 2, min_samples_leaf=2, max_features='sqrt',\\\n",
    "                                  max_depth = 20, bootstrap = True)\n",
    "forestbq.fit(Xb_train, yb_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 677,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8436118479221928"
      ]
     },
     "execution_count": 677,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forestbq.score(Xb_train, yb_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 678,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7022546419098143"
      ]
     },
     "execution_count": 678,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forestbq.score(xb_test, yb_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 679,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6647877984084881"
      ]
     },
     "execution_count": 679,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knnbq = KNeighborsClassifier(n_neighbors = 45)\n",
    "knnbq.fit(Xb_train, yb_train)\n",
    "knnbq.score(xb_test, yb_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 680,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5729442970822282"
      ]
     },
     "execution_count": 680,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnbbq = MultinomialNB()\n",
    "mnbbq.fit(Xb_train, yb_train)\n",
    "mnbbq.score(xb_test, yb_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 681,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/envs/metis/lib/python3.8/site-packages/sklearn/linear_model/_logistic.py:762: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  n_iter_i = _check_optimize_result(\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.6545092838196287"
      ]
     },
     "execution_count": 681,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logrbq = LogisticRegression()\n",
    "logrbq.fit(Xb_train_pre, yb_train)\n",
    "logrbq.score(xb_test_pre, yb_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 682,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7025862068965517"
      ]
     },
     "execution_count": 682,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbbq = xgb.XGBClassifier(n_estimators = 100)\n",
    "xgbbq.fit(Xb_train, yb_train)\n",
    "xgbbq.score(xb_test, yb_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 733,
   "metadata": {},
   "outputs": [
    {
     "ename": "KeyboardInterrupt",
     "evalue": "Interrupted by user",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-733-0d1fc5961644>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mstats\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_stats\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/Documents/ds/metis/repos/beer_project/get_stats.py\u001b[0m in \u001b[0;36mget_stats\u001b[0;34m()\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mget_stats\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m     \u001b[0mog\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfloat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Enter Original Gravity: \"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      6\u001b[0m     \u001b[0mfg\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfloat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Enter Final Gravity: \"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m     \u001b[0mabv\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfloat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Enter ABV: \"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/ipykernel/kernelbase.py\u001b[0m in \u001b[0;36mraw_input\u001b[0;34m(self, prompt)\u001b[0m\n\u001b[1;32m    858\u001b[0m                 \u001b[0;34m\"raw_input was called, but this frontend does not support input requests.\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    859\u001b[0m             )\n\u001b[0;32m--> 860\u001b[0;31m         return self._input_request(str(prompt),\n\u001b[0m\u001b[1;32m    861\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_parent_ident\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    862\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_parent_header\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/ipykernel/kernelbase.py\u001b[0m in \u001b[0;36m_input_request\u001b[0;34m(self, prompt, ident, parent, password)\u001b[0m\n\u001b[1;32m    902\u001b[0m             \u001b[0;32mexcept\u001b[0m \u001b[0mKeyboardInterrupt\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    903\u001b[0m                 \u001b[0;31m# re-raise KeyboardInterrupt, to truncate traceback\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 904\u001b[0;31m                 \u001b[0;32mraise\u001b[0m \u001b[0mKeyboardInterrupt\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Interrupted by user\"\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    905\u001b[0m             \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    906\u001b[0m                 \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlog\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwarning\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Invalid Message:\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexc_info\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: Interrupted by user"
     ]
    }
   ],
   "source": [
    "stats = get_stats()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 655,
   "metadata": {},
   "outputs": [],
   "source": [
    "from word_search import search_desc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 727,
   "metadata": {},
   "outputs": [
    {
     "name": "stdin",
     "output_type": "stream",
     "text": [
      "Enter beer review:   Enter beer review:  Hazy dank juicy stone fruit pine resin stone fruit mango citrus goodness for hoppy hop lovers. Lactose and hoppy bitter beer with award winning bitterness\n"
     ]
    }
   ],
   "source": [
    "review = input(\"Enter beer review: \")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 728,
   "metadata": {},
   "outputs": [],
   "source": [
    "desc = search_desc(review)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 714,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>dryh</th>\n",
       "      <th>lact</th>\n",
       "      <th>clone</th>\n",
       "      <th>juicy</th>\n",
       "      <th>dank</th>\n",
       "      <th>wedding</th>\n",
       "      <th>whisk</th>\n",
       "      <th>trop</th>\n",
       "      <th>award</th>\n",
       "      <th>best</th>\n",
       "      <th>...</th>\n",
       "      <th>dark</th>\n",
       "      <th>pine</th>\n",
       "      <th>sticky</th>\n",
       "      <th>foam</th>\n",
       "      <th>hoppy</th>\n",
       "      <th>bisc</th>\n",
       "      <th>ester</th>\n",
       "      <th>straw</th>\n",
       "      <th>choco</th>\n",
       "      <th>coffee</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1 rows × 36 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   dryh  lact  clone  juicy  dank  wedding  whisk  trop  award  best  ...  \\\n",
       "0     0     0      0      0     0        0      0     0      0     0  ...   \n",
       "\n",
       "   dark  pine  sticky  foam  hoppy  bisc  ester  straw  choco  coffee  \n",
       "0     0     0       0     0      0     0      0      0      0       0  \n",
       "\n",
       "[1 rows x 36 columns]"
      ]
     },
     "execution_count": 714,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "desc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 729,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>dryh</th>\n",
       "      <th>lact</th>\n",
       "      <th>clone</th>\n",
       "      <th>juicy</th>\n",
       "      <th>dank</th>\n",
       "      <th>wedding</th>\n",
       "      <th>whisk</th>\n",
       "      <th>trop</th>\n",
       "      <th>award</th>\n",
       "      <th>best</th>\n",
       "      <th>...</th>\n",
       "      <th>straw</th>\n",
       "      <th>choco</th>\n",
       "      <th>coffee</th>\n",
       "      <th>og</th>\n",
       "      <th>fg</th>\n",
       "      <th>abv</th>\n",
       "      <th>srm</th>\n",
       "      <th>ibu</th>\n",
       "      <th>bugu</th>\n",
       "      <th>hopcmp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.07</td>\n",
       "      <td>1.03</td>\n",
       "      <td>7.0</td>\n",
       "      <td>50.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>0.428571</td>\n",
       "      <td>20.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1 rows × 43 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   dryh  lact  clone  juicy  dank  wedding  whisk  trop  award  best  ...  \\\n",
       "0     0     1      0      1     1        0      0     1      1     0  ...   \n",
       "\n",
       "   straw  choco  coffee    og    fg  abv   srm   ibu      bugu  hopcmp  \n",
       "0      0      0       0  1.07  1.03  7.0  50.0  30.0  0.428571    20.0  \n",
       "\n",
       "[1 rows x 43 columns]"
      ]
     },
     "execution_count": 729,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bro_ = desc.join(stats)\n",
    "bro_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 716,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Strong/Dark Ales'], dtype=object)"
      ]
     },
     "execution_count": 716,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgbbq.predict(bro_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 730,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = [knnbq.predict(bro_)[0], mnbbq.predict(bro_)[0], logrbq.predict(bro_)[0],\\\n",
    "         forestbq.predict(bro_)[0], xgbbq.predict(bro_)[0], xgbbq.predict(bro_)[0]]\n",
    "clean_pred = defaultdict(int)\n",
    "for pred in preds:\n",
    "    if pred == 'IPA/Pale Ales':\n",
    "        clean_pred['IPA/Pale Ales'] +=1\n",
    "    elif pred == 'Strong/Dark Ales':\n",
    "        clean_pred['Strong/Dark Ales'] += 1\n",
    "    elif pred == 'Lager/Cream Ale':\n",
    "        clean_pred['Lager/Cream Ale'] += 1\n",
    "    else:\n",
    "        clean_pred['Other'] += 1\n",
    "res_dic = dict(clean_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 732,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({'Strong/Dark Ales': 5, 'IPA/Pale Ales': 1},\n",
       " ['Strong/Dark Ales',\n",
       "  'Strong/Dark Ales',\n",
       "  'IPA/Pale Ales',\n",
       "  'Strong/Dark Ales',\n",
       "  'Strong/Dark Ales',\n",
       "  'Strong/Dark Ales'])"
      ]
     },
     "execution_count": 732,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res_dic, preds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 731,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Strong/Dark Ales\n"
     ]
    }
   ],
   "source": [
    "for key, value in res_dic.items():\n",
    "    if value == len(preds)/2:\n",
    "        print(\"TIE\")\n",
    "        break\n",
    "    elif value > len(preds)//2:\n",
    "        print(key)\n",
    "    else:\n",
    "        continue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 702,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature ranking:\n",
      "1. feature 41 (0.211545)\n",
      "2. feature 39 (0.195041)\n",
      "3. feature 40 (0.165427)\n",
      "4. feature 42 (0.085557)\n",
      "5. feature 38 (0.071740)\n",
      "6. feature 26 (0.060191)\n",
      "7. feature 36 (0.059215)\n"
     ]
    },
    {
     "ename": "ValueError",
     "evalue": "shape mismatch: objects cannot be broadcast to a single shape",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-702-d433b3917b1d>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     13\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfigure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     14\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtitle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Feature importances\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 15\u001b[0;31m plt.bar(range(X.shape[1]), importances[indices],\n\u001b[0m\u001b[1;32m     16\u001b[0m         color=\"r\", yerr=std[indices], align=\"center\")\n\u001b[1;32m     17\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mxticks\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindices\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/matplotlib/pyplot.py\u001b[0m in \u001b[0;36mbar\u001b[0;34m(x, height, width, bottom, align, data, **kwargs)\u001b[0m\n\u001b[1;32m   2485\u001b[0m         \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mheight\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwidth\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0.8\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbottom\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malign\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'center'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2486\u001b[0m         data=None, **kwargs):\n\u001b[0;32m-> 2487\u001b[0;31m     return gca().bar(\n\u001b[0m\u001b[1;32m   2488\u001b[0m         \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mheight\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwidth\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mwidth\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbottom\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mbottom\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malign\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0malign\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2489\u001b[0m         **({\"data\": data} if data is not None else {}), **kwargs)\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/matplotlib/__init__.py\u001b[0m in \u001b[0;36minner\u001b[0;34m(ax, data, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1436\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0minner\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0max\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1437\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mdata\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1438\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0max\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0mmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msanitize_sequence\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1439\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1440\u001b[0m         \u001b[0mbound\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnew_sig\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbind\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0max\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/matplotlib/axes/_axes.py\u001b[0m in \u001b[0;36mbar\u001b[0;34m(self, x, height, width, bottom, align, **kwargs)\u001b[0m\n\u001b[1;32m   2428\u001b[0m                 \u001b[0myerr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_convert_dx\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0myerr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconvert_yunits\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2429\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2430\u001b[0;31m         x, height, width, y, linewidth = np.broadcast_arrays(\n\u001b[0m\u001b[1;32m   2431\u001b[0m             \u001b[0;31m# Make args iterable too.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2432\u001b[0m             np.atleast_1d(x), height, width, y, linewidth)\n",
      "\u001b[0;32m<__array_function__ internals>\u001b[0m in \u001b[0;36mbroadcast_arrays\u001b[0;34m(*args, **kwargs)\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/numpy/lib/stride_tricks.py\u001b[0m in \u001b[0;36mbroadcast_arrays\u001b[0;34m(subok, *args)\u001b[0m\n\u001b[1;32m    256\u001b[0m     \u001b[0margs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_m\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcopy\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msubok\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msubok\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0m_m\u001b[0m \u001b[0;32min\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    257\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 258\u001b[0;31m     \u001b[0mshape\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_broadcast_shape\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    259\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    260\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mshape\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0marray\u001b[0m \u001b[0;32min\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/numpy/lib/stride_tricks.py\u001b[0m in \u001b[0;36m_broadcast_shape\u001b[0;34m(*args)\u001b[0m\n\u001b[1;32m    187\u001b[0m     \u001b[0;31m# use the old-iterator because np.nditer does not handle size 0 arrays\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    188\u001b[0m     \u001b[0;31m# consistently\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 189\u001b[0;31m     \u001b[0mb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbroadcast\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m32\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    190\u001b[0m     \u001b[0;31m# unfortunately, it cannot handle 32 or more arguments directly\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    191\u001b[0m     \u001b[0;32mfor\u001b[0m \u001b[0mpos\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m32\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m31\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: shape mismatch: objects cannot be broadcast to a single shape"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "importances = forestbq.feature_importances_\n",
    "std = np.std([tree.feature_importances_ for tree in forestbq.estimators_],\n",
    "             axis=0)\n",
    "indices = np.argsort(importances)[::-1]\n",
    "\n",
    "# Print the feature ranking\n",
    "print(\"Feature ranking:\")\n",
    "\n",
    "for f in range(X.shape[1]):\n",
    "    print(\"%d. feature %d (%f)\" % (f + 1, indices[f], importances[indices[f]]))\n",
    "\n",
    "# Plot the impurity-based feature importances of the forest\n",
    "plt.figure()\n",
    "plt.title(\"Feature importances\")\n",
    "plt.bar(range(X.shape[1]), importances[indices],\n",
    "        color=\"r\", yerr=std[indices], align=\"center\")\n",
    "plt.xticks(range(X.shape[1]), indices)\n",
    "plt.xlim([-1, X.shape[1]])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 737,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1.062608801858081,\n",
       " 1.0145233461476855,\n",
       " 6.231271652573681,\n",
       " 13.537673704248506,\n",
       " 44.43336416786802,\n",
       " 35.467595707192054)"
      ]
     },
     "execution_count": 737,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.og.mean(), df.fg.mean(), df.abv.mean(), df.srm.mean(), df.ibu.mean(), df.hopcmp.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 739,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "No axis named 10 for object type Series",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/pandas/core/generic.py\u001b[0m in \u001b[0;36m_get_axis_number\u001b[0;34m(cls, axis)\u001b[0m\n\u001b[1;32m    367\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 368\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mcls\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_AXIS_TO_AXIS_NUMBER\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0maxis\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    369\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0mKeyError\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyError\u001b[0m: 10",
      "\nDuring handling of the above exception, another exception occurred:\n",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-739-5e6746d0f957>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mdf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mog\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/pandas/core/generic.py\u001b[0m in \u001b[0;36mstat_func\u001b[0;34m(self, axis, skipna, level, numeric_only, **kwargs)\u001b[0m\n\u001b[1;32m  11463\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mlevel\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m  11464\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_agg_by_level\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0maxis\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlevel\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlevel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mskipna\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mskipna\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m> 11465\u001b[0;31m         return self._reduce(\n\u001b[0m\u001b[1;32m  11466\u001b[0m             \u001b[0mfunc\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0maxis\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mskipna\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mskipna\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnumeric_only\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mnumeric_only\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m  11467\u001b[0m         )\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/pandas/core/series.py\u001b[0m in \u001b[0;36m_reduce\u001b[0;34m(self, op, name, axis, skipna, numeric_only, filter_type, **kwds)\u001b[0m\n\u001b[1;32m   4221\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   4222\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0maxis\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 4223\u001b[0;31m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_get_axis_number\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maxis\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   4224\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   4225\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdelegate\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mExtensionArray\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/anaconda3/envs/metis/lib/python3.8/site-packages/pandas/core/generic.py\u001b[0m in \u001b[0;36m_get_axis_number\u001b[0;34m(cls, axis)\u001b[0m\n\u001b[1;32m    368\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mcls\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_AXIS_TO_AXIS_NUMBER\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0maxis\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    369\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0mKeyError\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 370\u001b[0;31m             \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf\"No axis named {axis} for object type {cls.__name__}\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    371\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    372\u001b[0m     \u001b[0;34m@\u001b[0m\u001b[0mclassmethod\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: No axis named 10 for object type Series"
     ]
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:metis] *",
   "language": "python",
   "name": "conda-env-metis-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}