Word2Vec predictor with language.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df1 = pd.read_csv('./XYmat_lang_desc_cleaned.csv')\n",
    "df2 = pd.read_csv('./word_feature_mat.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>name</th>\n",
       "      <th>reviews</th>\n",
       "      <th>desc_snippet</th>\n",
       "      <th>FPS</th>\n",
       "      <th>Gore</th>\n",
       "      <th>Action</th>\n",
       "      <th>Demons</th>\n",
       "      <th>Shooter</th>\n",
       "      <th>First-Person</th>\n",
       "      <th>Great Soundtrack</th>\n",
       "      <th>...</th>\n",
       "      <th>Finnish</th>\n",
       "      <th>Norwegian</th>\n",
       "      <th>Swedish</th>\n",
       "      <th>Greek</th>\n",
       "      <th>Bulgarian</th>\n",
       "      <th>Romanian</th>\n",
       "      <th>Unnamed: 412</th>\n",
       "      <th>Vietnamese</th>\n",
       "      <th>Slovakian</th>\n",
       "      <th>number of languages</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DOOM</td>\n",
       "      <td>92</td>\n",
       "      <td>Now includes all three premium DLC packs (Unto...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</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>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PLAYERUNKNOWN'S BATTLEGROUNDS</td>\n",
       "      <td>49</td>\n",
       "      <td>PLAYERUNKNOWN'S BATTLEGROUNDS is a battle roya...</td>\n",
       "      <td>1</td>\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>...</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>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BATTLETECH</td>\n",
       "      <td>71</td>\n",
       "      <td>Take command of your own mercenary outfit of '...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</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>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>DayZ</td>\n",
       "      <td>61</td>\n",
       "      <td>The post-soviet country of Chernarus is struck...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</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>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>EVE Online</td>\n",
       "      <td>74</td>\n",
       "      <td>EVE Online is a community-driven spaceship MMO...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</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>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 416 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                            name  reviews  \\\n",
       "0                           DOOM       92   \n",
       "1  PLAYERUNKNOWN'S BATTLEGROUNDS       49   \n",
       "2                     BATTLETECH       71   \n",
       "3                           DayZ       61   \n",
       "4                     EVE Online       74   \n",
       "\n",
       "                                        desc_snippet  FPS  Gore  Action  \\\n",
       "0  Now includes all three premium DLC packs (Unto...    1     1       1   \n",
       "1  PLAYERUNKNOWN'S BATTLEGROUNDS is a battle roya...    1     0       1   \n",
       "2  Take command of your own mercenary outfit of '...    0     0       1   \n",
       "3  The post-soviet country of Chernarus is struck...    1     0       1   \n",
       "4  EVE Online is a community-driven spaceship MMO...    0     0       1   \n",
       "\n",
       "   Demons  Shooter  First-Person  Great Soundtrack  ...  Finnish  Norwegian  \\\n",
       "0       1        1             1                 1  ...        0          0   \n",
       "1       0        1             1                 0  ...        0          0   \n",
       "2       0        0             0                 1  ...        0          0   \n",
       "3       0        1             0                 0  ...        0          0   \n",
       "4       0        0             0                 0  ...        0          0   \n",
       "\n",
       "   Swedish  Greek  Bulgarian  Romanian  Unnamed: 412  Vietnamese  Slovakian  \\\n",
       "0        0      0          0         0             0           0          0   \n",
       "1        0      0          0         0             0           0          0   \n",
       "2        0      0          0         0             0           0          0   \n",
       "3        0      0          0         0             0           0          0   \n",
       "4        0      0          0         0             0           0          0   \n",
       "\n",
       "   number of languages  \n",
       "0                   10  \n",
       "1                   17  \n",
       "2                    4  \n",
       "3                    9  \n",
       "4                    4  \n",
       "\n",
       "[5 rows x 416 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>name</th>\n",
       "      <th>desc_snippet</th>\n",
       "      <th>wf1</th>\n",
       "      <th>wf2</th>\n",
       "      <th>wf3</th>\n",
       "      <th>wf4</th>\n",
       "      <th>wf5</th>\n",
       "      <th>wf6</th>\n",
       "      <th>wf7</th>\n",
       "      <th>wf8</th>\n",
       "      <th>...</th>\n",
       "      <th>wf91</th>\n",
       "      <th>wf92</th>\n",
       "      <th>wf93</th>\n",
       "      <th>wf94</th>\n",
       "      <th>wf95</th>\n",
       "      <th>wf96</th>\n",
       "      <th>wf97</th>\n",
       "      <th>wf98</th>\n",
       "      <th>wf99</th>\n",
       "      <th>wf100</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DOOM</td>\n",
       "      <td>Now includes all three premium DLC packs (Unto...</td>\n",
       "      <td>0.006291</td>\n",
       "      <td>0.019894</td>\n",
       "      <td>-0.036758</td>\n",
       "      <td>-0.083445</td>\n",
       "      <td>0.005177</td>\n",
       "      <td>0.061250</td>\n",
       "      <td>-0.021474</td>\n",
       "      <td>0.017422</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.022482</td>\n",
       "      <td>-0.033148</td>\n",
       "      <td>-0.068780</td>\n",
       "      <td>-0.001817</td>\n",
       "      <td>-0.017904</td>\n",
       "      <td>0.030351</td>\n",
       "      <td>0.004494</td>\n",
       "      <td>0.032382</td>\n",
       "      <td>-0.038966</td>\n",
       "      <td>-0.011268</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PLAYERUNKNOWN'S BATTLEGROUNDS</td>\n",
       "      <td>PLAYERUNKNOWN'S BATTLEGROUNDS is a battle roya...</td>\n",
       "      <td>0.065965</td>\n",
       "      <td>-0.040907</td>\n",
       "      <td>0.083006</td>\n",
       "      <td>-0.127466</td>\n",
       "      <td>0.047808</td>\n",
       "      <td>-0.001543</td>\n",
       "      <td>-0.050009</td>\n",
       "      <td>0.039216</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.023967</td>\n",
       "      <td>0.092510</td>\n",
       "      <td>-0.061511</td>\n",
       "      <td>-0.053102</td>\n",
       "      <td>-0.076502</td>\n",
       "      <td>0.028666</td>\n",
       "      <td>-0.070288</td>\n",
       "      <td>0.085814</td>\n",
       "      <td>-0.036982</td>\n",
       "      <td>0.007031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BATTLETECH</td>\n",
       "      <td>Take command of your own mercenary outfit of '...</td>\n",
       "      <td>-0.031684</td>\n",
       "      <td>-0.044110</td>\n",
       "      <td>0.039126</td>\n",
       "      <td>-0.083009</td>\n",
       "      <td>0.019074</td>\n",
       "      <td>-0.035856</td>\n",
       "      <td>-0.076101</td>\n",
       "      <td>-0.022778</td>\n",
       "      <td>...</td>\n",
       "      <td>0.018407</td>\n",
       "      <td>0.000874</td>\n",
       "      <td>-0.036874</td>\n",
       "      <td>0.007654</td>\n",
       "      <td>-0.051571</td>\n",
       "      <td>0.053892</td>\n",
       "      <td>-0.045746</td>\n",
       "      <td>0.104181</td>\n",
       "      <td>-0.009850</td>\n",
       "      <td>0.006609</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>DayZ</td>\n",
       "      <td>The post-soviet country of Chernarus is struck...</td>\n",
       "      <td>-0.055274</td>\n",
       "      <td>-0.039872</td>\n",
       "      <td>0.011899</td>\n",
       "      <td>-0.009593</td>\n",
       "      <td>-0.051783</td>\n",
       "      <td>-0.071177</td>\n",
       "      <td>-0.109749</td>\n",
       "      <td>-0.064325</td>\n",
       "      <td>...</td>\n",
       "      <td>0.005817</td>\n",
       "      <td>0.016707</td>\n",
       "      <td>-0.070429</td>\n",
       "      <td>0.028132</td>\n",
       "      <td>-0.015154</td>\n",
       "      <td>0.082653</td>\n",
       "      <td>-0.020251</td>\n",
       "      <td>0.084262</td>\n",
       "      <td>0.054496</td>\n",
       "      <td>-0.026706</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>EVE Online</td>\n",
       "      <td>EVE Online is a community-driven spaceship MMO...</td>\n",
       "      <td>0.043309</td>\n",
       "      <td>0.011002</td>\n",
       "      <td>0.063850</td>\n",
       "      <td>-0.116439</td>\n",
       "      <td>0.026295</td>\n",
       "      <td>-0.004743</td>\n",
       "      <td>-0.111624</td>\n",
       "      <td>-0.012195</td>\n",
       "      <td>...</td>\n",
       "      <td>0.030270</td>\n",
       "      <td>0.049835</td>\n",
       "      <td>-0.096783</td>\n",
       "      <td>-0.020521</td>\n",
       "      <td>-0.073409</td>\n",
       "      <td>0.023094</td>\n",
       "      <td>-0.035156</td>\n",
       "      <td>0.090076</td>\n",
       "      <td>-0.001013</td>\n",
       "      <td>-0.026774</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 102 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                            name  \\\n",
       "0                           DOOM   \n",
       "1  PLAYERUNKNOWN'S BATTLEGROUNDS   \n",
       "2                     BATTLETECH   \n",
       "3                           DayZ   \n",
       "4                     EVE Online   \n",
       "\n",
       "                                        desc_snippet       wf1       wf2  \\\n",
       "0  Now includes all three premium DLC packs (Unto...  0.006291  0.019894   \n",
       "1  PLAYERUNKNOWN'S BATTLEGROUNDS is a battle roya...  0.065965 -0.040907   \n",
       "2  Take command of your own mercenary outfit of '... -0.031684 -0.044110   \n",
       "3  The post-soviet country of Chernarus is struck... -0.055274 -0.039872   \n",
       "4  EVE Online is a community-driven spaceship MMO...  0.043309  0.011002   \n",
       "\n",
       "        wf3       wf4       wf5       wf6       wf7       wf8  ...      wf91  \\\n",
       "0 -0.036758 -0.083445  0.005177  0.061250 -0.021474  0.017422  ... -0.022482   \n",
       "1  0.083006 -0.127466  0.047808 -0.001543 -0.050009  0.039216  ... -0.023967   \n",
       "2  0.039126 -0.083009  0.019074 -0.035856 -0.076101 -0.022778  ...  0.018407   \n",
       "3  0.011899 -0.009593 -0.051783 -0.071177 -0.109749 -0.064325  ...  0.005817   \n",
       "4  0.063850 -0.116439  0.026295 -0.004743 -0.111624 -0.012195  ...  0.030270   \n",
       "\n",
       "       wf92      wf93      wf94      wf95      wf96      wf97      wf98  \\\n",
       "0 -0.033148 -0.068780 -0.001817 -0.017904  0.030351  0.004494  0.032382   \n",
       "1  0.092510 -0.061511 -0.053102 -0.076502  0.028666 -0.070288  0.085814   \n",
       "2  0.000874 -0.036874  0.007654 -0.051571  0.053892 -0.045746  0.104181   \n",
       "3  0.016707 -0.070429  0.028132 -0.015154  0.082653 -0.020251  0.084262   \n",
       "4  0.049835 -0.096783 -0.020521 -0.073409  0.023094 -0.035156  0.090076   \n",
       "\n",
       "       wf99     wf100  \n",
       "0 -0.038966 -0.011268  \n",
       "1 -0.036982  0.007031  \n",
       "2 -0.009850  0.006609  \n",
       "3  0.054496 -0.026706  \n",
       "4 -0.001013 -0.026774  \n",
       "\n",
       "[5 rows x 102 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df2.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "array1 = df1.to_numpy()\n",
    "X1 = np.delete(array1, 2, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[['DOOM' 92 1 ... 0 0 10]\n",
      " [\"PLAYERUNKNOWN'S BATTLEGROUNDS\" 49 1 ... 0 0 17]\n",
      " ['BATTLETECH' 71 0 ... 0 0 4]\n",
      " ...\n",
      " ['The Last Weekend' 57 0 ... 0 0 1]\n",
      " ['Zulu Response' 47 1 ... 0 0 2]\n",
      " ['Choice of Rebels: Uprising' 96 0 ... 0 0 6]]\n"
     ]
    }
   ],
   "source": [
    "print(X1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "array2 = df2.to_numpy()\n",
    "X2 = np.delete(array2, 1, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[['DOOM' 0.006291399244219065 0.019894296675920485 ...\n",
      "  0.032381537929177286 -0.03896602783352136 -0.01126769632101059]\n",
      " [\"PLAYERUNKNOWN'S BATTLEGROUNDS\" 0.06596502251923084 -0.0409067090600729\n",
      "  ... 0.08581372052431106 -0.03698160322383046 0.007031131908297539]\n",
      " ['BATTLETECH' -0.03168355114758015 -0.044109889492392536 ...\n",
      "  0.10418071886524558 -0.009849852696061137 0.006608905643224717]\n",
      " ...\n",
      " ['Galactis' 0.029017702862620356 -0.02262181304395199 ...\n",
      "  0.041712296288460494 0.010209517553448677 0.0016313848667778073]\n",
      " ['Alive' -0.02112476322799921 -0.025149831059388814 ...\n",
      "  0.059446010022657 0.030122669786214824 -0.04876134330406785]\n",
      " ['Stories In Stone' -0.0131360721308738 0.01890635378658772 ...\n",
      "  0.02485982086509466 0.04401859268546104 -0.0465050233528018]]\n"
     ]
    }
   ],
   "source": [
    "print(X2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10277 24612\n",
      "(10203, 514)\n"
     ]
    }
   ],
   "source": [
    "new_array = []\n",
    "i = 0\n",
    "\n",
    "while i < len(X1):\n",
    "    for arr in X2:\n",
    "        if X1[i][0] == arr[0]:\n",
    "            new_array.append(np.append(X1[i][1:], arr[1:]))\n",
    "            break\n",
    "    i += 1\n",
    "    #print(i)\n",
    "\n",
    "print(len(X1), len(X2))\n",
    "\n",
    "Xy = np.asarray(new_array)\n",
    "print(Xy.shape)\n",
    "for i in Xy:\n",
    "    for j in i:\n",
    "        if np.isnan(j) == True:\n",
    "            print(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[92 1 1 ... 0.032381537929177286 -0.03896602783352136\n",
      "  -0.01126769632101059]\n",
      " [49 1 0 ... 0.08581372052431106 -0.03698160322383046\n",
      "  0.007031131908297539]\n",
      " [71 0 0 ... 0.10418071886524558 -0.009849852696061137\n",
      "  0.006608905643224717]\n",
      " ...\n",
      " [57 0 0 ... 0.039271814282983535 0.03507877103984356 0.00484479907900095]\n",
      " [47 1 0 ... 0.08566247075796128 -0.05540984738618136\n",
      "  -0.004844885074999183]\n",
      " [96 0 0 ... 0.0850942999124527 -0.006312704086303711\n",
      "  -0.014720835583284497]]\n"
     ]
    }
   ],
   "source": [
    "print(Xy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = Xy[:, 1:]\n",
    "y = Xy[:, 0]\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Split train test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(7140, 513) (3063, 513) (7140,) (3063,)\n",
      "(array([], dtype=int64), array([], dtype=int64))\n"
     ]
    }
   ],
   "source": [
    "num_train = len(Xy) // 10 * 7\n",
    "\n",
    "Xtrain = X[:num_train]\n",
    "Xtest = X[num_train:]\n",
    "ytrain = y[:num_train]\n",
    "ytest = y[num_train:]\n",
    "print(Xtrain.shape, Xtest.shape, ytrain.shape, ytest.shape)\n",
    "print(np.where(Xy=='nan'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test data with multiple linear regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "import scipy as sp\n",
    "import scipy.sparse.linalg\n",
    "import math"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(513,)\n"
     ]
    }
   ],
   "source": [
    "out = sp.sparse.linalg.lsqr(Xtrain,ytrain)\n",
    "beta = out[0]\n",
    "print(beta.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compute Baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "multiple variable loss=2.24e+01\n"
     ]
    }
   ],
   "source": [
    "yBaseline = np.ones(len(ytest)) * np.mean(ytrain)\n",
    "errs = np.array(abs(ytest - yBaseline))\n",
    "loss = np.sum(errs)\n",
    "avgL = loss/len(ytest)\n",
    "print(\"multiple variable loss=\"+\"{:.2e}\".format(avgL))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Compute loss with linear regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0 0 1 ... -0.012236664863303305 -0.0187341008335352\n",
      "  0.012150854244828223]\n",
      " [0 0 1 ... 0.029344024229794742 0.03323996104300022 -0.03156232042238116]\n",
      " [0 0 0 ... 0.00985959842801094 0.03081499636173248 0.009103235136717559]\n",
      " ...\n",
      " [0 0 1 ... 0.039271814282983535 0.03507877103984356 0.00484479907900095]\n",
      " [1 0 0 ... 0.08566247075796128 -0.05540984738618136\n",
      "  -0.004844885074999183]\n",
      " [0 0 0 ... 0.0850942999124527 -0.006312704086303711\n",
      "  -0.014720835583284497]]\n",
      "multiple variable loss=2.03e+01\n"
     ]
    }
   ],
   "source": [
    "print(Xtest)\n",
    "yp = Xtest@beta\n",
    "errs = np.array(abs(ytest - yp))\n",
    "avgL = np.mean(errs)\n",
    "\n",
    "\n",
    "print(\"multiple variable loss=\"+\"{:.2e}\".format(avgL))\n",
    "\n",
    "arr = []\n",
    "for i in range(len(ytest)):\n",
    "    arr.append(abs(yp[i]-ytest[i]))\n",
    "arr.sort(reverse = True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'int'>\n"
     ]
    }
   ],
   "source": [
    "print(type(Xtest[-1][2]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "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",
    "\n",
    "plt.plot(yp, label = \"yp\")\n",
    "plt.plot(ytest, label = \"ytest\")\n",
    "plt.legend(loc=\"upper left\")\n",
    "axes = plt.gca()\n",
    "#axes.set_xlim([0,500])\n",
    "#axes.set_ylim([0,150])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x133200790>]"
      ]
     },
     "execution_count": 22,
     "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": [
    "plt.plot(errs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20.338622799882195\n"
     ]
    }
   ],
   "source": [
    "print(np.mean(errs))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}