2-4 군집화 K-Means.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. K-means 알고리즘 \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "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>sepal_length</th>\n",
       "      <th>sepal_width</th>\n",
       "      <th>petal_length</th>\n",
       "      <th>petal_width</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5.1</td>\n",
       "      <td>3.5</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4.9</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.7</td>\n",
       "      <td>3.2</td>\n",
       "      <td>1.3</td>\n",
       "      <td>0.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   sepal_length  sepal_width  petal_length  petal_width\n",
       "0           5.1          3.5           1.4          0.2\n",
       "1           4.9          3.0           1.4          0.2\n",
       "2           4.7          3.2           1.3          0.2"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import scale\n",
    "from sklearn.datasets import load_iris\n",
    "from sklearn.cluster import KMeans\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "%matplotlib inline\n",
    "\n",
    "iris=load_iris()\n",
    "irisDF=pd.DataFrame(data=iris.data,columns=['sepal_length','sepal_width','petal_length','petal_width'])\n",
    "irisDF.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 붓꽃 데이터 세트를 3개의 그룹으로 군집화"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KMeans(algorithm='auto', copy_x=True, init='k-means++', max_iter=300,\n",
       "    n_clusters=3, n_init=10, n_jobs=None, precompute_distances='auto',\n",
       "    random_state=0, tol=0.0001, verbose=0)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kmeans=KMeans(n_clusters=3, init='k-means++',max_iter=300, random_state=0)\n",
    "kmeans.fit(irisDF)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
      " 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      " 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 2 2 2 2 0 2 2 2 2\n",
      " 2 2 0 0 2 2 2 2 0 2 0 2 0 2 2 0 0 2 2 2 2 2 0 2 2 2 2 0 2 2 2 0 2 2 2 0 2\n",
      " 2 0]\n"
     ]
    }
   ],
   "source": [
    "print(kmeans.labels_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "target  cluster\n",
      "0       1          50\n",
      "1       0          48\n",
      "        2           2\n",
      "2       0          14\n",
      "        2          36\n",
      "Name: sepal_length, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "irisDF['target']=iris.target\n",
    "irisDF['cluster']=kmeans.labels_\n",
    "\n",
    "# target값과 cluster값 개수 비교\n",
    "iris_result=irisDF.groupby(['target','cluster'])['sepal_length'].count()\n",
    "print(iris_result)\n",
    "\n",
    "# target 0 = cluster 1\n",
    "# 1 = 0, 2 = 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 군집화 시각화하기\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>sepal_length</th>\n",
       "      <th>sepal_width</th>\n",
       "      <th>petal_length</th>\n",
       "      <th>petal_width</th>\n",
       "      <th>target</th>\n",
       "      <th>cluster</th>\n",
       "      <th>pca_x</th>\n",
       "      <th>pca_y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5.1</td>\n",
       "      <td>3.5</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>-2.684126</td>\n",
       "      <td>0.319397</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4.9</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>-2.714142</td>\n",
       "      <td>-0.177001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.7</td>\n",
       "      <td>3.2</td>\n",
       "      <td>1.3</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>-2.888991</td>\n",
       "      <td>-0.144949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.6</td>\n",
       "      <td>3.1</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>-2.745343</td>\n",
       "      <td>-0.318299</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5.0</td>\n",
       "      <td>3.6</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>-2.728717</td>\n",
       "      <td>0.326755</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   sepal_length  sepal_width  petal_length  petal_width  target  cluster  \\\n",
       "0           5.1          3.5           1.4          0.2       0        1   \n",
       "1           4.9          3.0           1.4          0.2       0        1   \n",
       "2           4.7          3.2           1.3          0.2       0        1   \n",
       "3           4.6          3.1           1.5          0.2       0        1   \n",
       "4           5.0          3.6           1.4          0.2       0        1   \n",
       "\n",
       "      pca_x     pca_y  \n",
       "0 -2.684126  0.319397  \n",
       "1 -2.714142 -0.177001  \n",
       "2 -2.888991 -0.144949  \n",
       "3 -2.745343 -0.318299  \n",
       "4 -2.728717  0.326755  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.decomposition import PCA\n",
    "# 4개의 속성을 2개로 차원 축소\n",
    "pca=PCA(n_components=2)\n",
    "pca_transformed = pca.fit_transform(iris.data)\n",
    "\n",
    "# x, y 좌표로 개별 데이터를 표현\n",
    "irisDF['pca_x']=pca_transformed[:,0]\n",
    "irisDF['pca_y']=pca_transformed[:,1]\n",
    "irisDF.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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": [
    "# 각 클러스터에 해당하는 index 추출\n",
    "marker0_ind = irisDF[irisDF['cluster']==0].index\n",
    "marker1_ind = irisDF[irisDF['cluster']==1].index\n",
    "marker2_ind = irisDF[irisDF['cluster']==2].index\n",
    "\n",
    "# 각 클러스터에 해당하는 index로 각 cluster레벨의 pca_x, pca_y값 추출\n",
    "plt.scatter(x=irisDF.loc[marker0_ind,'pca_x'],y=irisDF.loc[marker0_ind,'pca_y'],marker='o')\n",
    "plt.scatter(x=irisDF.loc[marker1_ind,'pca_x'],y=irisDF.loc[marker1_ind,'pca_y'],marker='s')\n",
    "plt.scatter(x=irisDF.loc[marker2_ind,'pca_x'],y=irisDF.loc[marker2_ind,'pca_y'],marker='^')\n",
    "\n",
    "plt.xlabel('PCA 1')\n",
    "plt.ylabel('PCA 2')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1- 4. 군집화 알고리즘 테스트를 위한 데이터 생성\n",
    "#### 1) make_blobs()\n",
    "- make_blobs()로 만든 피처 데이터 세트의 군집화 분포 확인"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(200, 2) (200,)\n",
      "[0 1 2] [67 67 66]\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn.datasets import make_blobs\n",
    "%matplotlib inline\n",
    "\n",
    "X, y=make_blobs(n_samples=200, n_features=2, centers=3, cluster_std=0.8, random_state=0)\n",
    "print(X.shape, y.shape)\n",
    "\n",
    "# y target 값의 분포를 확인\n",
    "unique, counts = np.unique(y, return_counts=True)\n",
    "print(unique, counts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>ftr1</th>\n",
       "      <th>ftr2</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-1.692427</td>\n",
       "      <td>3.622025</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.697940</td>\n",
       "      <td>4.428867</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.100228</td>\n",
       "      <td>4.606317</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       ftr1      ftr2  target\n",
       "0 -1.692427  3.622025       2\n",
       "1  0.697940  4.428867       0\n",
       "2  1.100228  4.606317       0"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "clusterDF=pd.DataFrame(data=X, columns=['ftr1','ftr2'])\n",
    "clusterDF['target']=y\n",
    "clusterDF.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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": [
    "target_list = np.unique(y)  # array([0, 1, 2])\n",
    "\n",
    "markers=['o', 's','^','P','D','H','X']\n",
    "\n",
    "# target별로 marker 지정하고, scatter plot 그리기\n",
    "for target in target_list:\n",
    "    target_cluster = clusterDF[clusterDF['target']==target]\n",
    "    plt.scatter(x=target_cluster['ftr1'],y=target_cluster['ftr2'],edgecolor='k',marker=markers[target])\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Kmeans 군집화를 수행한 뒤 군집별로 시작화 하기."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "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>ftr1</th>\n",
       "      <th>ftr2</th>\n",
       "      <th>target</th>\n",
       "      <th>kmeans_label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-1.692427</td>\n",
       "      <td>3.622025</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.697940</td>\n",
       "      <td>4.428867</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.100228</td>\n",
       "      <td>4.606317</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       ftr1      ftr2  target  kmeans_label\n",
       "0 -1.692427  3.622025       2             1\n",
       "1  0.697940  4.428867       0             0\n",
       "2  1.100228  4.606317       0             0"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# X데이터 k-means 클러스터링 수행\n",
    "kmeans = KMeans(n_clusters=3, init = 'k-means++', max_iter=200, random_state=0)\n",
    "cluster_labels = kmeans.fit_predict(X)\n",
    "clusterDF['kmeans_label'] = cluster_labels\n",
    "clusterDF.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "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": [
    "# 각 클러스터의 중심 위치 추출 (중심위치 좌표 시각화를 위해서)\n",
    "centers = kmeans.cluster_centers_\n",
    "unique_labels = np.unique(cluster_labels)\n",
    "markers=['o','s','^','P','D','H','x']\n",
    "\n",
    "# label 유형별로  반복하면서\n",
    "for label in unique_labels:\n",
    "    # scatter plot 그리기\n",
    "    label_cluster = clusterDF[clusterDF['kmeans_label']==label]\n",
    "    center_x_y = centers[label]\n",
    "    plt.scatter(x=label_cluster['ftr1'], y=label_cluster['ftr2'],edgecolor='k',marker=markers[label])\n",
    "    \n",
    "    # 군집별 중심 위치 좌표 시각화\n",
    "    plt.scatter(x=center_x_y[0],y=center_x_y[1],s=200,color='white',\n",
    "               alpha=0.9, edgecolor='k', marker=markers[label])\n",
    "    plt.scatter(x=center_x_y[0],y=center_x_y[1],s=70,color='k',edgecolor='k',\n",
    "               marker='$%d$' % label)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "target  kmeans_label\n",
      "0       0               66\n",
      "        1                1\n",
      "1       2               67\n",
      "2       1               65\n",
      "        2                1\n",
      "Name: kmeans_label, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "# target과 cluster 매핑 결과 확인\n",
    "# 대부분 잘 매핑된 것을 볼 수 있다.\n",
    "\n",
    "print(clusterDF.groupby('target')['kmeans_label'].value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- make_blobs()의 cluster_std 파라미터로 데이터의 분포도 조절하기.\n",
    "* cluster_std 가 작을수록 군집 중심에 데이터가 모여 있고, 클수록 데이터가 퍼져있다"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. 군집평가 \n",
    "### 4. 실루엣 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "scrolled": 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>sepal_length</th>\n",
       "      <th>sepal_width</th>\n",
       "      <th>petal_length</th>\n",
       "      <th>petal_width</th>\n",
       "      <th>cluster</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5.1</td>\n",
       "      <td>3.5</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4.9</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.7</td>\n",
       "      <td>3.2</td>\n",
       "      <td>1.3</td>\n",
       "      <td>0.2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   sepal_length  sepal_width  petal_length  petal_width  cluster\n",
       "0           5.1          3.5           1.4          0.2        1\n",
       "1           4.9          3.0           1.4          0.2        1\n",
       "2           4.7          3.2           1.3          0.2        1"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import scale\n",
    "from sklearn.datasets import load_iris\n",
    "from sklearn.cluster import KMeans\n",
    "\n",
    "# 실루엣 분석 metrc값을 구하기 위한 API 추가\n",
    "from sklearn.metrics import silhouette_samples, silhouette_score\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "%matplotlib inline\n",
    "\n",
    "iris=load_iris()\n",
    "feature_names=['sepal_length','sepal_width','petal_length','petal_width']\n",
    "irisDF=pd.DataFrame(data=iris.data,columns=feature_names)\n",
    "kmeans=KMeans(n_clusters=3, init='k-means++',max_iter=300, random_state=0).fit(irisDF)\n",
    "\n",
    "irisDF['cluster']=kmeans.labels_\n",
    "irisDF.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Silhouette_sample(데이터 셋, 군집 레이블) : 개별 데이터의 실루엣 계수\n",
    "- silhouette_score: 모든 데이터의 실루엣 계수값을 평균"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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>sepal_length</th>\n",
       "      <th>sepal_width</th>\n",
       "      <th>petal_length</th>\n",
       "      <th>petal_width</th>\n",
       "      <th>cluster</th>\n",
       "      <th>silhouette_coeff</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5.1</td>\n",
       "      <td>3.5</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.2</td>\n",
       "      <td>1</td>\n",
       "      <td>0.852955</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4.9</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.4</td>\n",
       "      <td>0.2</td>\n",
       "      <td>1</td>\n",
       "      <td>0.815495</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.7</td>\n",
       "      <td>3.2</td>\n",
       "      <td>1.3</td>\n",
       "      <td>0.2</td>\n",
       "      <td>1</td>\n",
       "      <td>0.829315</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   sepal_length  sepal_width  petal_length  petal_width  cluster  \\\n",
       "0           5.1          3.5           1.4          0.2        1   \n",
       "1           4.9          3.0           1.4          0.2        1   \n",
       "2           4.7          3.2           1.3          0.2        1   \n",
       "\n",
       "   silhouette_coeff  \n",
       "0          0.852955  \n",
       "1          0.815495  \n",
       "2          0.829315  "
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# iris의 모든 개별 데이터에 실루엣 계수값을 구함\n",
    "score_samples=silhouette_samples(iris.data,irisDF['cluster'])\n",
    "\n",
    "irisDF['silhouette_coeff']=score_samples\n",
    "irisDF.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "붓꽃 데이터셋 Silhouette Analysis Score:0.553\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "cluster\n",
       "0    0.417320\n",
       "1    0.798140\n",
       "2    0.451105\n",
       "Name: silhouette_coeff, dtype: float64"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# silhouette_score: 모든 데이터의 실루엣 계수값을 평균\n",
    "average_score=silhouette_score(iris.data, irisDF['cluster'])\n",
    "print('붓꽃 데이터셋 Silhouette Analysis Score:{0:.3f}'.format(average_score))\n",
    "\n",
    "# 군집별 실루엣 스코어 평균\n",
    "irisDF.groupby('cluster')['silhouette_coeff'].mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 좋은 군집화는 Silhouette_score 값이 0 ~ 1 사이의 값을 갖고, 1에 가까울수록 좋음.\n",
    "- 하지만 전체 실루엣 계수의 평균값과 더불어 개별 군집의 평균값의 편차가 크지 않아야 함"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 실루엣 시각화\n",
    "- visualize_silhouette 함수: 군집 개수를 변화시키면서 K-평균 군집을 수행했을 때 \n",
    "    개별 군집별 평균 실루엣 계수 값을 시각화해서 군집의 개수를 정하는 데 도움을 줌\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "### 여러개의 클러스터링 갯수를 List로 입력 받아 각각의 실루엣 계수를 면적으로 시각화한 함수 작성\n",
    "def visualize_silhouette(cluster_lists, X_features): \n",
    "    \n",
    "    from sklearn.datasets import make_blobs\n",
    "    from sklearn.cluster import KMeans\n",
    "    from sklearn.metrics import silhouette_samples, silhouette_score\n",
    "\n",
    "    import matplotlib.pyplot as plt\n",
    "    import matplotlib.cm as cm\n",
    "    import math\n",
    "    \n",
    "    # 입력값으로 클러스터링 갯수들을 리스트로 받아서, 각 갯수별로 클러스터링을 적용하고 실루엣 개수를 구함\n",
    "    n_cols = len(cluster_lists)\n",
    "    \n",
    "    # plt.subplots()으로 리스트에 기재된 클러스터링 수만큼의 sub figures를 가지는 axs 생성 \n",
    "    fig, axs = plt.subplots(figsize=(4*n_cols, 4), nrows=1, ncols=n_cols)\n",
    "    \n",
    "    # 리스트의 클러스터링 갯수들을 차례로 iteration 수행하면서 실루엣 개수 시각화\n",
    "    for ind, n_cluster in enumerate(cluster_lists):\n",
    "        \n",
    "        # KMeans 클러스터링 수행\n",
    "        clusterer = KMeans(n_clusters = n_cluster, max_iter=500, random_state=0)\n",
    "        cluster_labels = clusterer.fit_predict(X_features)\n",
    "        \n",
    "        # 실루엣 스코어 (평균)\n",
    "        sil_avg = silhouette_score(X_features, cluster_labels)\n",
    "        # 모든 개별 데이터 실루엣 계수값 \n",
    "        sil_values = silhouette_samples(X_features, cluster_labels)\n",
    "        \n",
    "        y_lower = 10\n",
    "        axs[ind].set_title('Number of Cluster : '+ str(n_cluster)+'\\n' \\\n",
    "                          'Silhouette Score :' + str(round(sil_avg,3)) )\n",
    "        axs[ind].set_xlabel(\"The silhouette coefficient values\")\n",
    "        axs[ind].set_ylabel(\"Cluster label\")\n",
    "        axs[ind].set_xlim([-0.1, 1])\n",
    "        axs[ind].set_ylim([0, len(X_features) + (n_cluster + 1) * 10])\n",
    "        axs[ind].set_yticks([])  # Clear the yaxis labels / ticks\n",
    "        axs[ind].set_xticks([0, 0.2, 0.4, 0.6, 0.8, 1])\n",
    "        \n",
    "        # 클러스터링 갯수별로 fill_betweenx( )형태의 막대 그래프 표현. \n",
    "        for i in range(n_cluster):\n",
    "            ith_cluster_sil_values = sil_values[cluster_labels==i]\n",
    "            ith_cluster_sil_values.sort()\n",
    "            \n",
    "            size_cluster_i = ith_cluster_sil_values.shape[0]\n",
    "            y_upper = y_lower + size_cluster_i\n",
    "            \n",
    "            color = cm.nipy_spectral(float(i) / n_cluster)\n",
    "            axs[ind].fill_betweenx(np.arange(y_lower, y_upper), 0, ith_cluster_sil_values, \\\n",
    "                                facecolor=color, edgecolor=color, alpha=0.7)\n",
    "            axs[ind].text(-0.05, y_lower + 0.5 * size_cluster_i, str(i))\n",
    "            y_lower = y_upper + 10\n",
    "            \n",
    "        axs[ind].axvline(x=sil_avg, color=\"red\", linestyle=\"--\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x288 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# make_blobs 을 통해 clustering 을 위한 4개의 클러스터 중심의 500개 2차원 데이터 셋 생성  \n",
    "from sklearn.datasets import make_blobs\n",
    "X, y = make_blobs(n_samples=500, n_features=2, centers=4, cluster_std=1, \\\n",
    "                  center_box=(-10.0, 10.0), shuffle=True, random_state=1)   # center_box: center값을 -10~10 사이로 지정\n",
    "\n",
    "# cluster 개수를 2개, 3개, 4개, 5개 일때의 클러스터별 실루엣 계수 평균값을 시각화 \n",
    "visualize_silhouette([ 2, 3, 4, 5], X)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4개가 가장 이상적인 군집 개수\n",
    "\n",
    "- cluster 개수가 2개일 경우: 0번 군집에서 반 이상의 데이터가 평균을 넘지 못함.\n",
    "- cluster 3 : 2번 군집에서 대부분의 데이터가 평균을 넘지 못함.\n",
    "- cluster 4 : 평균 실루엣 계수 값이 작지만 개별 군집의 평균 실루엣 계수 값이 비교적 균일하게 위치\n",
    "\n",
    "\n",
    "Second best를 고려하기\n",
    "- 일반적으로 2개의 군집으로 나눴을 때, 평균 실루엣(실루엣 스코어)가 제일 높게 나온다.\n",
    "- 하지만 이는 적절하지 않을 때가 많으므로, Second best를 고려해서 선택하는 것이 좋다. (이 경우에는 4개)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 군집평가 실습  (붓꽃 데이터)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "wss=[]\n",
    "sis=[]\n",
    "\n",
    "for k in range(2,11):\n",
    "    kmeans=KMeans(n_clusters=k, random_state=0)\n",
    "    kmeans.fit(iris.data)\n",
    "    wss=np.append(wss, kmeans.inertia_)\n",
    "    sis=np.append(sis, silhouette_score(iris.data, kmeans.labels_))\n",
    "\n",
    "flg,ax1=plt.subplots()\n",
    "ax2=ax1.twinx()\n",
    "\n",
    "line1=ax1.plot(np.arange(2,11),wss,'ro--',label='SSE')\n",
    "line2=ax2.plot(np.arange(2,11),sis,'bo--',label='Silhouette')\n",
    "\n",
    "ax1.set_ylim(wss.min()*0.95, wss.max()*1.05)\n",
    "ax2.set_ylim(sis.min()*0.95, sis.max()*1.05)\n",
    "\n",
    "ax1.set_xlabel('군집 수')\n",
    "ax1.set_ylabel('SSE')\n",
    "ax2.set_ylabel('Silhouette score')\n",
    "\n",
    "lines=line1+line2\n",
    "labels=[l.get_label() for l in lines]\n",
    "\n",
    "plt.legend(lines, labels)\n",
    "plt.show()\n",
    "\n",
    "\n",
    "# SSE와 Silhouette score 그래프를 바탕으로 Elbow 기법을 활용하면\n",
    "# 붓꽃 데이터의 최적 군집 범위는 2, 3, 4, 5로 정함"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x288 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.datasets import load_iris\n",
    "iris=load_iris()\n",
    "visualize_silhouette([2,3,4,5], iris.data)\n",
    "\n",
    "# 군집 개수가 2개일 때 가장 최적. "
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}