Model III -- Clustering & Anomaly Detection -- COPD Included.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import statsmodels.formula.api as smf\n",
    "import matplotlib.pylab as plt\n",
    "import math\n",
    "from scipy import stats\n",
    "import geopandas as gpd\n",
    "%matplotlib inline\n",
    "from sklearn import preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "Model_III = pd.read_csv('Model III Delta 08-13.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>Neighborhood-index</th>\n",
       "      <th>Adults Reporting Exercise in the Past 30 Days</th>\n",
       "      <th>Adults with Asthma in the Past 12 Months</th>\n",
       "      <th>Adults with High Cholesterol</th>\n",
       "      <th>Adults without a Smoke-Free Home Policy</th>\n",
       "      <th>Asthma Hospitalizations (Adults)</th>\n",
       "      <th>Asthma Hospitalizations (Children 0 to 4 Yrs Old)</th>\n",
       "      <th>Asthma Hospitalizations (Children 5 to 17 Yrs Old)</th>\n",
       "      <th>Cigarette Smoking among Adults</th>\n",
       "      <th>Chronic Obstructive Pulmonary Disease  Hospitalization</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Bayside - Little Neck</td>\n",
       "      <td>0.033333</td>\n",
       "      <td>0.365854</td>\n",
       "      <td>0.180495</td>\n",
       "      <td>-0.096774</td>\n",
       "      <td>0.081081</td>\n",
       "      <td>0.543224</td>\n",
       "      <td>0.607735</td>\n",
       "      <td>-0.171642</td>\n",
       "      <td>0.025947</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Bedford Stuyvesant - Crown Heights</td>\n",
       "      <td>0.058901</td>\n",
       "      <td>-0.890625</td>\n",
       "      <td>0.230132</td>\n",
       "      <td>0.372937</td>\n",
       "      <td>0.078453</td>\n",
       "      <td>0.042582</td>\n",
       "      <td>-0.009199</td>\n",
       "      <td>0.290476</td>\n",
       "      <td>-0.217956</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Bensonhurst - Bay Ridge</td>\n",
       "      <td>-0.064801</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.051702</td>\n",
       "      <td>0.061303</td>\n",
       "      <td>-0.193673</td>\n",
       "      <td>0.165584</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.077670</td>\n",
       "      <td>-0.058596</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Borough Park</td>\n",
       "      <td>0.071749</td>\n",
       "      <td>-0.042857</td>\n",
       "      <td>0.207116</td>\n",
       "      <td>0.011111</td>\n",
       "      <td>0.210544</td>\n",
       "      <td>0.109612</td>\n",
       "      <td>0.067039</td>\n",
       "      <td>-0.798246</td>\n",
       "      <td>0.033032</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Canarsie - Flatlands</td>\n",
       "      <td>0.180593</td>\n",
       "      <td>-0.416667</td>\n",
       "      <td>0.184300</td>\n",
       "      <td>0.125000</td>\n",
       "      <td>0.062060</td>\n",
       "      <td>-0.264089</td>\n",
       "      <td>-0.495238</td>\n",
       "      <td>0.390244</td>\n",
       "      <td>-0.124864</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   Neighborhood-index  \\\n",
       "0               Bayside - Little Neck   \n",
       "1  Bedford Stuyvesant - Crown Heights   \n",
       "2             Bensonhurst - Bay Ridge   \n",
       "3                        Borough Park   \n",
       "4                Canarsie - Flatlands   \n",
       "\n",
       "   Adults Reporting Exercise in the Past 30 Days  \\\n",
       "0                                       0.033333   \n",
       "1                                       0.058901   \n",
       "2                                      -0.064801   \n",
       "3                                       0.071749   \n",
       "4                                       0.180593   \n",
       "\n",
       "   Adults with Asthma in the Past 12 Months  Adults with High Cholesterol  \\\n",
       "0                                  0.365854                      0.180495   \n",
       "1                                 -0.890625                      0.230132   \n",
       "2                                  0.000000                      0.051702   \n",
       "3                                 -0.042857                      0.207116   \n",
       "4                                 -0.416667                      0.184300   \n",
       "\n",
       "   Adults without a Smoke-Free Home Policy  Asthma Hospitalizations (Adults)  \\\n",
       "0                                -0.096774                          0.081081   \n",
       "1                                 0.372937                          0.078453   \n",
       "2                                 0.061303                         -0.193673   \n",
       "3                                 0.011111                          0.210544   \n",
       "4                                 0.125000                          0.062060   \n",
       "\n",
       "   Asthma Hospitalizations (Children 0 to 4 Yrs Old)  \\\n",
       "0                                           0.543224   \n",
       "1                                           0.042582   \n",
       "2                                           0.165584   \n",
       "3                                           0.109612   \n",
       "4                                          -0.264089   \n",
       "\n",
       "   Asthma Hospitalizations (Children 5 to 17 Yrs Old)  \\\n",
       "0                                           0.607735    \n",
       "1                                          -0.009199    \n",
       "2                                           0.000000    \n",
       "3                                           0.067039    \n",
       "4                                          -0.495238    \n",
       "\n",
       "   Cigarette Smoking among Adults  \\\n",
       "0                       -0.171642   \n",
       "1                        0.290476   \n",
       "2                        0.077670   \n",
       "3                       -0.798246   \n",
       "4                        0.390244   \n",
       "\n",
       "   Chronic Obstructive Pulmonary Disease  Hospitalization  \n",
       "0                                           0.025947       \n",
       "1                                          -0.217956       \n",
       "2                                          -0.058596       \n",
       "3                                           0.033032       \n",
       "4                                          -0.124864       "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Model_III.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[]\n"
     ]
    }
   ],
   "source": [
    "from pgmpy.estimators import HillClimbSearch\n",
    "from pgmpy.estimators import BicScore\n",
    "hc = HillClimbSearch(Model_III, scoring_method=BicScore(Model_III))\n",
    "best_model = hc.estimate()\n",
    "print(best_model.edges())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(Model_III.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = Model_III.iloc[:,1:11]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Gaussian Mixture Clustering -- Model III"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GaussianMixture(covariance_type='full', init_params='kmeans', max_iter=100,\n",
       "        means_init=None, n_components=3, n_init=1, precisions_init=None,\n",
       "        random_state=999, reg_covar=1e-06, tol=0.001, verbose=0,\n",
       "        verbose_interval=10, warm_start=False, weights_init=None)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.mixture import GaussianMixture\n",
    "GM=GaussianMixture(n_components=3,random_state=999)\n",
    "GM.fit(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# score_samples function gives the log of the probability density of each data record given its cluster.\n",
    "# Note that probability densities can exceed 1 (unlike probabilities of discrete data).\n",
    "import matplotlib.pylab as plt\n",
    "plt.figure(figsize=(20,4))\n",
    "plt.scatter(range(len(X)),GM.score_samples(X)) #gives each data record's log-likelihood. allows for direct anamoly detection\n",
    "plt.plot(GM.score_samples(X),\"r--\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>-1.636656</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.402578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.625390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.792031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.833108</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           0\n",
       "34 -1.636656\n",
       "27  0.402578\n",
       "31  0.625390\n",
       "21  0.792031\n",
       "22  0.833108"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 5 most anomalous data points by log-likelihood (i.e. log probability density)\n",
    "pd.DataFrame(GM.score_samples(X)).sort_values(0)[:5]"
   ]
  },
  {
   "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>Neighborhood-index</th>\n",
       "      <th>Chronic Obstructive Pulmonary Disease  Hospitalization</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>Sunset Park</td>\n",
       "      <td>-0.122460</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Port Richmond</td>\n",
       "      <td>-0.786517</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>Southeast Queens</td>\n",
       "      <td>0.120125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Jamaica</td>\n",
       "      <td>0.133152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Kingsbridge - Riverdale</td>\n",
       "      <td>0.390357</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Neighborhood-index  \\\n",
       "34              Sunset Park   \n",
       "27            Port Richmond   \n",
       "31         Southeast Queens   \n",
       "21                  Jamaica   \n",
       "22  Kingsbridge - Riverdale   \n",
       "\n",
       "    Chronic Obstructive Pulmonary Disease  Hospitalization  \n",
       "34                                          -0.122460       \n",
       "27                                          -0.786517       \n",
       "31                                           0.120125       \n",
       "21                                           0.133152       \n",
       "22                                           0.390357       "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Neighborhood_COPD = Model_III[['Neighborhood-index', 'Chronic Obstructive Pulmonary Disease  Hospitalization']]\n",
    "Neighborhood_COPD.iloc[[34,27, 31, 21, 22],:]\n",
    "\n",
    "#Model_I[6]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## K-means Model III"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_scaled = preprocessing.scale(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2.097196</td>\n",
       "      <td>3.351571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4.318163</td>\n",
       "      <td>2.855510</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3.099059</td>\n",
       "      <td>2.307283</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.915323</td>\n",
       "      <td>3.324488</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5.208994</td>\n",
       "      <td>3.510582</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          0         1\n",
       "0  2.097196  3.351571\n",
       "1  4.318163  2.855510\n",
       "2  3.099059  2.307283\n",
       "3  2.915323  3.324488\n",
       "4  5.208994  3.510582"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "KM=KMeans(n_clusters=2,random_state=999)\n",
    "scor=KM.fit_predict(X_scaled)\n",
    "\n",
    "# distance to each cluster center\n",
    "res=pd.DataFrame(KM.transform(X_scaled))\n",
    "res.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>cluster</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>5.916428</td>\n",
       "      <td>4.339659</td>\n",
       "      <td>1</td>\n",
       "      <td>4.339659</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>4.780823</td>\n",
       "      <td>4.280376</td>\n",
       "      <td>1</td>\n",
       "      <td>4.280376</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>3.754415</td>\n",
       "      <td>5.102781</td>\n",
       "      <td>0</td>\n",
       "      <td>3.754415</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>3.726292</td>\n",
       "      <td>4.528731</td>\n",
       "      <td>0</td>\n",
       "      <td>3.726292</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>4.063064</td>\n",
       "      <td>3.645671</td>\n",
       "      <td>1</td>\n",
       "      <td>3.645671</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           0         1  cluster     score\n",
       "27  5.916428  4.339659        1  4.339659\n",
       "41  4.780823  4.280376        1  4.280376\n",
       "18  3.754415  5.102781        0  3.754415\n",
       "31  3.726292  4.528731        0  3.726292\n",
       "34  4.063064  3.645671        1  3.645671"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res=pd.concat((res,pd.DataFrame(KM.fit_predict(X_scaled))),axis=1)\n",
    "res.columns=list(range(2))+[\"cluster\"]\n",
    "res.loc[:,\"score\"]=res.apply(lambda x: x[int(x[\"cluster\"])],axis=1)\n",
    "\n",
    "# find data records farthest from cluster centers\n",
    "res.sort_values(\"score\",ascending=False)[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>Neighborhood-index</th>\n",
       "      <th>Chronic Obstructive Pulmonary Disease  Hospitalization</th>\n",
       "      <th>K-means</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Port Richmond</td>\n",
       "      <td>-0.786517</td>\n",
       "      <td>4.339659</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>Willowbrook</td>\n",
       "      <td>-0.103976</td>\n",
       "      <td>4.280376</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Greenwich Village - SoHo</td>\n",
       "      <td>0.225806</td>\n",
       "      <td>3.754415</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>Southeast Queens</td>\n",
       "      <td>0.120125</td>\n",
       "      <td>3.726292</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>Sunset Park</td>\n",
       "      <td>-0.122460</td>\n",
       "      <td>3.645671</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Neighborhood-index  \\\n",
       "27             Port Richmond   \n",
       "41               Willowbrook   \n",
       "18  Greenwich Village - SoHo   \n",
       "31          Southeast Queens   \n",
       "34               Sunset Park   \n",
       "\n",
       "    Chronic Obstructive Pulmonary Disease  Hospitalization   K-means  \n",
       "27                                          -0.786517       4.339659  \n",
       "41                                          -0.103976       4.280376  \n",
       "18                                           0.225806       3.754415  \n",
       "31                                           0.120125       3.726292  \n",
       "34                                          -0.122460       3.645671  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = res.loc[:,\"score\"]\n",
    "df_N = Model_III\n",
    "df_N = df_N.reset_index()\n",
    "df_N['K-means'] = x\n",
    "df_K = df_N[['Neighborhood-index', 'Chronic Obstructive Pulmonary Disease  Hospitalization', 'K-means']]\n",
    "df_K.nlargest(5, 'K-means')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = Model_III.iloc[:,1:3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "y['cluster'] = (res.loc[:,\"cluster\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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": [
    "#plt.subplot(1,2,2)\n",
    "plt.scatter(y.iloc[:,0], y.iloc[:,1], c=y['cluster'], cmap=plt.cm.cool)\n",
    "#plt.xlabel('x1')\n",
    "#plt.ylabel('x2')\n",
    "#plt.xticks(())\n",
    "#plt.yticks(())\n",
    "plt.title(\"Kmeans-{} clusters\".format(2))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "For n_clusters =2, the average silhouette_score is :0.59476526546\n",
      "For n_clusters =3, the average silhouette_score is :0.488766692282\n",
      "For n_clusters =4, the average silhouette_score is :0.574409503523\n",
      "For n_clusters =5, the average silhouette_score is :0.570337468513\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import silhouette_samples, silhouette_score\n",
    "import matplotlib.cm as cm\n",
    "import numpy as np\n",
    "\n",
    "range_n_clusters = [2, 3, 4,5]\n",
    "for n_clusters in range_n_clusters: #try different numbers of clusters\n",
    "    km = KMeans(n_clusters=n_clusters, random_state=999)\n",
    "    cluster_labels = km.fit_predict(y.iloc[:,0:2])\n",
    "    #report average Silhouette score\n",
    "    silhouette_avg = silhouette_score(y.iloc[:,0:2], cluster_labels)\n",
    "    print(\"For n_clusters ={},\".format(n_clusters)+\" the average silhouette_score is :{}\".format(silhouette_avg))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Isolation Forest Modeling -- Model III"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import IsolationForest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\aniha\\Anaconda3\\envs\\TwentySevenVersion\\lib\\site-packages\\sklearn\\ensemble\\iforest.py:213: FutureWarning: default contamination parameter 0.1 will change in version 0.22 to \"auto\". This will change the predict method behavior.\n",
      "  FutureWarning)\n",
      "C:\\Users\\aniha\\Anaconda3\\envs\\TwentySevenVersion\\lib\\site-packages\\sklearn\\ensemble\\iforest.py:223: FutureWarning: behaviour=\"old\" is deprecated and will be removed in version 0.22. Please use behaviour=\"new\", which makes the decision_function change to match other anomaly detection algorithm API.\n",
      "  FutureWarning)\n",
      "C:\\Users\\aniha\\Anaconda3\\envs\\TwentySevenVersion\\lib\\site-packages\\sklearn\\ensemble\\iforest.py:250: UserWarning: max_samples (50) is greater than the total number of samples (42). max_samples will be set to n_samples for estimation.\n",
      "  % (self.max_samples, n_samples))\n",
      "C:\\Users\\aniha\\Anaconda3\\envs\\TwentySevenVersion\\lib\\site-packages\\sklearn\\ensemble\\iforest.py:417: DeprecationWarning: threshold_ attribute is deprecated in 0.20 and will be removed in 0.22.\n",
      "  \" be removed in 0.22.\", DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "clf = IsolationForest(max_samples=50, random_state=123)\n",
    "clf.fit_predict(X)\n",
    "IsoForest = (clf.score_samples(X))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "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>Neighborhood-index</th>\n",
       "      <th>IsoFoScore</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Port Richmond</td>\n",
       "      <td>-0.519062</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>Willowbrook</td>\n",
       "      <td>-0.507593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Canarsie - Flatlands</td>\n",
       "      <td>-0.492987</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Flushing - Clearview</td>\n",
       "      <td>-0.484400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>East Flatbush - Flatbush</td>\n",
       "      <td>-0.477965</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Neighborhood-index  IsoFoScore\n",
       "27             Port Richmond   -0.519062\n",
       "41               Willowbrook   -0.507593\n",
       "4       Canarsie - Flatlands   -0.492987\n",
       "13      Flushing - Clearview   -0.484400\n",
       "10  East Flatbush - Flatbush   -0.477965"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Iso_score = GM.score_samples(Data_Numeric.iloc[:,1:32])\n",
    "df_I = Model_III\n",
    "df_I = df_I.reset_index()\n",
    "df_I['IsoFoScore'] = IsoForest\n",
    "df_I = df_I[['Neighborhood-index','IsoFoScore']]\n",
    "df_I.nsmallest(5, 'IsoFoScore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LOF Clustering -- Model III"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\aniha\\Anaconda3\\envs\\TwentySevenVersion\\lib\\site-packages\\sklearn\\neighbors\\lof.py:236: FutureWarning: default contamination parameter 0.1 will change in version 0.22 to \"auto\". This will change the predict method behavior.\n",
      "  FutureWarning)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.neighbors import LocalOutlierFactor\n",
    "\n",
    "# fit the model\n",
    "clf_LOF = LocalOutlierFactor(n_neighbors=5)\n",
    "y_pred = clf_LOF.fit_predict(X)\n",
    "LOF = (clf_LOF.negative_outlier_factor_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>Neighborhood-index</th>\n",
       "      <th>Chronic Obstructive Pulmonary Disease  Hospitalization</th>\n",
       "      <th>LOF_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Port Richmond</td>\n",
       "      <td>-0.786517</td>\n",
       "      <td>-1.738410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>Willowbrook</td>\n",
       "      <td>-0.103976</td>\n",
       "      <td>-1.510757</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>Sunset Park</td>\n",
       "      <td>-0.122460</td>\n",
       "      <td>-1.327384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Greenwich Village - SoHo</td>\n",
       "      <td>0.225806</td>\n",
       "      <td>-1.293548</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Gramercy Park - Murray Hill</td>\n",
       "      <td>-0.003549</td>\n",
       "      <td>-1.232164</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Neighborhood-index  \\\n",
       "27                Port Richmond   \n",
       "41                  Willowbrook   \n",
       "34                  Sunset Park   \n",
       "18     Greenwich Village - SoHo   \n",
       "16  Gramercy Park - Murray Hill   \n",
       "\n",
       "    Chronic Obstructive Pulmonary Disease  Hospitalization  LOF_Score  \n",
       "27                                          -0.786517       -1.738410  \n",
       "41                                          -0.103976       -1.510757  \n",
       "34                                          -0.122460       -1.327384  \n",
       "18                                           0.225806       -1.293548  \n",
       "16                                          -0.003549       -1.232164  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fit the model\n",
    "df_LOF = Model_III\n",
    "df_LOF = df_LOF.reset_index()\n",
    "df_LOF['LOF_Score'] = LOF\n",
    "df_LOF = df_LOF[['Neighborhood-index', 'Chronic Obstructive Pulmonary Disease  Hospitalization', 'LOF_Score']]\n",
    "df_LOF.nsmallest(5, 'LOF_Score')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_reshape = X.values.reshape(-2,2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "189"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(X_reshape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# fit the model\n",
    "clf_LOF = LocalOutlierFactor(n_neighbors=7)\n",
    "y_pred = clf_LOF.fit_predict(X_reshape)\n",
    "#y_pred_outliers = y_pred[200:]\n",
    "\n",
    "# plot the level sets of the decision function\n",
    "xx, yy = np.meshgrid(np.linspace(-10, 10, 100), np.linspace(-10, 10, 100))\n",
    "Z = clf_LOF._decision_function(np.c_[xx.ravel(), yy.ravel()])\n",
    "Z = Z.reshape(xx.shape)\n",
    "\n",
    "plt.figure(figsize=(10,10))\n",
    "\n",
    "\n",
    "plt.title(\"Local Outlier Factor (LOF)\")\n",
    "plt.contourf(xx, yy, Z, cmap=plt.cm.Blues_r)\n",
    "\n",
    "a = plt.scatter(X_reshape[:189, 0], X_reshape[:189, 1], c='black')\n",
    "b = plt.scatter(X_reshape[189:, 0], X_reshape[189:, 1], c='red')\n",
    "plt.axis('tight')\n",
    "plt.xlim((-5, 5))\n",
    "plt.ylim((-5, 5))\n",
    "plt.legend([a, b],\n",
    "           [\"normal observations\",\n",
    "            \"abnormal observations\"],\n",
    "           loc=\"upper left\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DBSCAN -- Model I"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.cluster import DBSCAN\n",
    "from sklearn.datasets.samples_generator import make_blobs\n",
    "import matplotlib.pyplot as plt\n",
    "import sklearn.datasets as data\n",
    "import sklearn.cluster as cluster\n",
    "import time\n",
    "from sklearn import cluster, datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "def visuDB(data,eps,min_samples):\n",
    "    db = DBSCAN(eps, min_samples).fit(X_reshape)\n",
    "    plt.figure(figsize=(6,6))\n",
    "    plt.scatter(X_reshape[:,0],X_reshape[:,1],c=db.labels_,s=10)\n",
    "    print \"Proportion of unclustered points=\",(1*(db.labels_ < 0)).mean()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Proportion of unclustered points= 0.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "visuDB(X_reshape,0.5,7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "# your code here\n",
    "\n",
    "total_num_samples = len(X_reshape[:,0])\n",
    "max_eps = max(np.max(X_reshape[:,0])-np.min(X_reshape[:,0]),np.max(X_reshape[:,1])-np.min(X_reshape[:,1]))\n",
    "\n",
    "for eps in np.logspace(np.log10(max_eps)-3,np.log10(max_eps),10):\n",
    "    for min_samples in np.linspace(40,800,40,endpoint=True):\n",
    "        db = DBSCAN(eps, min_samples).fit(X_reshape)\n",
    "        num_clustered = len(db.labels_[db.labels_ >= 0])\n",
    "        if (num_clustered > 0) & (num_clustered < 0.1*total_num_samples):\n",
    "            dbb=db\n",
    "            plt.figure(figsize=(8,8))\n",
    "            print \"eps =\", eps, \", min_samples =\", min_samples, \", num_detected_points =\", len(db.labels_[db.labels_ >= 0])\n",
    "            \n",
    "            # evaluation\n",
    "            num_correct_clustered = len(db.labels_[0:100][db.labels_[0:100] >= 0])\n",
    "            print \"precision=\",1.0*num_correct_clustered/num_clustered,\"recall=\",1.0*num_correct_clustered/100.0\n",
    "            \n",
    "            plt.scatter(X_reshape[:,0],X_reshape[:,1],c=db.labels_,s=10,cmap=plt.cm.cool)\n",
    "            plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}