Untitled.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Crime category prediction by description"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is sample of multiple class Text Classification with Scikit-Learn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import necessary base librarys\n",
    "import os\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from io import StringIO"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "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>Unnamed: 0</th>\n",
       "      <th>Date</th>\n",
       "      <th>Category</th>\n",
       "      <th>Description</th>\n",
       "      <th>Day of Week</th>\n",
       "      <th>PdDistrict</th>\n",
       "      <th>Resolution</th>\n",
       "      <th>Address</th>\n",
       "      <th>Longitude</th>\n",
       "      <th>Latitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1.580712e+09</td>\n",
       "      <td>Missing Person</td>\n",
       "      <td>Found Person</td>\n",
       "      <td>Monday</td>\n",
       "      <td>Taraval</td>\n",
       "      <td>Open or Active</td>\n",
       "      <td>20TH AVE \\ WINSTON DR</td>\n",
       "      <td>-122.476039</td>\n",
       "      <td>37.726950</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1.580673e+09</td>\n",
       "      <td>Stolen Property</td>\n",
       "      <td>Stolen Property, Possession with Knowledge, Re...</td>\n",
       "      <td>Monday</td>\n",
       "      <td>Mission</td>\n",
       "      <td>Cite or Arrest Adult</td>\n",
       "      <td>24TH ST \\ SHOTWELL ST</td>\n",
       "      <td>-122.415172</td>\n",
       "      <td>37.752440</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>6</td>\n",
       "      <td>1.579104e+09</td>\n",
       "      <td>Offences Against The Family And Children</td>\n",
       "      <td>Elder Adult or Dependent Abuse (not Embezzleme...</td>\n",
       "      <td>Thursday</td>\n",
       "      <td>Mission</td>\n",
       "      <td>Unfounded</td>\n",
       "      <td>18TH ST \\ SANCHEZ ST</td>\n",
       "      <td>-122.430574</td>\n",
       "      <td>37.761155</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>11</td>\n",
       "      <td>1.578647e+09</td>\n",
       "      <td>Lost Property</td>\n",
       "      <td>Lost Property</td>\n",
       "      <td>Friday</td>\n",
       "      <td>Park</td>\n",
       "      <td>Open or Active</td>\n",
       "      <td>VISTA LN \\ CRESTLINE DR</td>\n",
       "      <td>-122.446245</td>\n",
       "      <td>37.751669</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>12</td>\n",
       "      <td>1.580728e+09</td>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft, From Person, $200-$950 (other than Pick...</td>\n",
       "      <td>Monday</td>\n",
       "      <td>Bayview</td>\n",
       "      <td>Open or Active</td>\n",
       "      <td>03RD ST \\ KIRKWOOD AVE</td>\n",
       "      <td>-122.389587</td>\n",
       "      <td>37.737955</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0          Date                                  Category  \\\n",
       "0           0  1.580712e+09                            Missing Person   \n",
       "1           1  1.580673e+09                           Stolen Property   \n",
       "2           6  1.579104e+09  Offences Against The Family And Children   \n",
       "3          11  1.578647e+09                             Lost Property   \n",
       "4          12  1.580728e+09                             Larceny Theft   \n",
       "\n",
       "                                         Description Day of Week PdDistrict  \\\n",
       "0                                       Found Person      Monday    Taraval   \n",
       "1  Stolen Property, Possession with Knowledge, Re...      Monday    Mission   \n",
       "2  Elder Adult or Dependent Abuse (not Embezzleme...    Thursday    Mission   \n",
       "3                                      Lost Property      Friday       Park   \n",
       "4  Theft, From Person, $200-$950 (other than Pick...      Monday    Bayview   \n",
       "\n",
       "             Resolution                  Address   Longitude   Latitude  \n",
       "0        Open or Active    20TH AVE \\ WINSTON DR -122.476039  37.726950  \n",
       "1  Cite or Arrest Adult    24TH ST \\ SHOTWELL ST -122.415172  37.752440  \n",
       "2             Unfounded     18TH ST \\ SANCHEZ ST -122.430574  37.761155  \n",
       "3        Open or Active  VISTA LN \\ CRESTLINE DR -122.446245  37.751669  \n",
       "4        Open or Active   03RD ST \\ KIRKWOOD AVE -122.389587  37.737955  "
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Read data from file and create first dataframe\n",
    "df = pd.read_csv(\"train.csv\", sep='\\t')\n",
    "# Preview the first 5 lines of the loaded data \n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "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>Category</th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Missing Person</td>\n",
       "      <td>Found Person</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Stolen Property</td>\n",
       "      <td>Stolen Property, Possession with Knowledge, Re...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Offences Against The Family And Children</td>\n",
       "      <td>Elder Adult or Dependent Abuse (not Embezzleme...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Lost Property</td>\n",
       "      <td>Lost Property</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft, From Person, $200-$950 (other than Pick...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                   Category  \\\n",
       "0                            Missing Person   \n",
       "1                           Stolen Property   \n",
       "2  Offences Against The Family And Children   \n",
       "3                             Lost Property   \n",
       "4                             Larceny Theft   \n",
       "\n",
       "                                         Description  \n",
       "0                                       Found Person  \n",
       "1  Stolen Property, Possession with Knowledge, Re...  \n",
       "2  Elder Adult or Dependent Abuse (not Embezzleme...  \n",
       "3                                      Lost Property  \n",
       "4  Theft, From Person, $200-$950 (other than Pick...  "
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Select columns Category & Description and make it as new dataframe\n",
    "df1 = df[['Category','Description']]\n",
    "df1.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 218438 entries, 0 to 218437\n",
      "Data columns (total 2 columns):\n",
      "Category       218438 non-null object\n",
      "Description    218438 non-null object\n",
      "dtypes: object(2)\n",
      "memory usage: 3.3+ MB\n"
     ]
    }
   ],
   "source": [
    "# Check df1 dataframe info\n",
    "df1.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Arson', 'Assault', 'Burglary', 'Civil Sidewalks', 'Courtesy Report', 'Disorderly Conduct', 'Drug Offense', 'Embezzlement', 'Family Offense', 'Fire Report', 'Forgery And Counterfeiting', 'Fraud', 'Larceny Theft', 'Lost Property', 'Missing Person', 'Motor Vehicle Theft', 'Offences Against The Family And Children', 'Prostitution', 'Recovered Vehicle', 'Robbery', 'Sex Offense', 'Stolen Property', 'Suspicious Occ', 'Traffic Collision', 'Traffic Violation Arrest', 'Vandalism', 'Vehicle Impounded', 'Warrant', 'Weapons Carrying Etc', 'Weapons Offense']\n"
     ]
    }
   ],
   "source": [
    "# List out all category type in Category\n",
    "labels = list(df1.Category.unique())\n",
    "labels.sort()\n",
    "print(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Arson', 'Assault', 'Burglary', 'Civil Sidewalks', 'Courtesy Report', 'Disorderly Conduct', 'Drug Offense', 'Embezzlement', 'Family Offense', 'Fire Report', 'Forgery And Counterfeiting', 'Fraud', 'Larceny Theft', 'Lost Property', 'Missing Person', 'Motor Vehicle Theft', 'Offences Against The Family And Children', 'Prostitution', 'Recovered Vehicle', 'Robbery', 'Sex Offense', 'Stolen Property', 'Suspicious Occ', 'Traffic Collision', 'Traffic Violation Arrest', 'Vandalism', 'Vehicle Impounded', 'Warrant', 'Weapons Carrying Etc', 'Weapons Offense']\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot bar graph to show data distribution of category type in Category\n",
    "labels = list(df.Category.unique())\n",
    "labels.sort()\n",
    "print(labels)\n",
    "plt.figure(figsize=(14,6))\n",
    "df1.Category.value_counts().plot(kind='bar')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    },
    {
     "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>Category</th>\n",
       "      <th>Description</th>\n",
       "      <th>category_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Missing Person</td>\n",
       "      <td>Found Person</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Stolen Property</td>\n",
       "      <td>Stolen Property, Possession with Knowledge, Re...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Offences Against The Family And Children</td>\n",
       "      <td>Elder Adult or Dependent Abuse (not Embezzleme...</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Lost Property</td>\n",
       "      <td>Lost Property</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft, From Person, $200-$950 (other than Pick...</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                   Category  \\\n",
       "0                            Missing Person   \n",
       "1                           Stolen Property   \n",
       "2  Offences Against The Family And Children   \n",
       "3                             Lost Property   \n",
       "4                             Larceny Theft   \n",
       "\n",
       "                                         Description  category_id  \n",
       "0                                       Found Person            0  \n",
       "1  Stolen Property, Possession with Knowledge, Re...            1  \n",
       "2  Elder Adult or Dependent Abuse (not Embezzleme...            2  \n",
       "3                                      Lost Property            3  \n",
       "4  Theft, From Person, $200-$950 (other than Pick...            4  "
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Factorize category type in Category and map it into new category_id column\n",
    "df1['category_id'] = df1['Category'].factorize()[0]\n",
    "category_id_df1 = df1[['Category', 'category_id']].drop_duplicates().sort_values('category_id')\n",
    "category_to_id = dict(category_id_df1.values)\n",
    "id_to_category = dict(category_id_df1[['category_id', 'Category']].values)\n",
    "df1.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot bar graph to show data distribution of category id in category_id\n",
    "fig = plt.figure(figsize=(8,6))\n",
    "df1.groupby('category_id').Description.count().plot.bar(ylim=0)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(218438, 1211)"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Extract text from Description \n",
    "# https://scikit-learn.org/stable/modules/generated/sklearn.feature_extraction.text.TfidfVectorizer.html\n",
    "# Now, each of 218438 Description is represented by 1211 features, representing the tf-idf score for different unigrams and bigrams.\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "tfidf = TfidfVectorizer(sublinear_tf=True, min_df=5, norm='l2', encoding='latin-1', ngram_range=(1, 2), stop_words='english')\n",
    "features = tfidf.fit_transform(df1.Description).toarray()\n",
    "labels = df1.category_id\n",
    "features.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# 'Arson':\n",
      "  . Most correlated unigrams:\n",
      ". commercial\n",
      ". arson\n",
      "  . Most correlated bigrams:\n",
      ". arson inhabited\n",
      ". arson vehicle\n",
      "# 'Assault':\n",
      "  . Most correlated unigrams:\n",
      ". aggravated\n",
      ". battery\n",
      "  . Most correlated bigrams:\n",
      ". aggravated weapon\n",
      ". assault aggravated\n",
      "# 'Burglary':\n",
      "  . Most correlated unigrams:\n",
      ". burglary\n",
      ". entry\n",
      "  . Most correlated bigrams:\n",
      ". forcible entry\n",
      ". unlawful entry\n",
      "# 'Civil Sidewalks':\n",
      "  . Most correlated unigrams:\n",
      ". civil\n",
      ". sidewalks\n",
      "  . Most correlated bigrams:\n",
      ". sidewalks citation\n",
      ". civil sidewalks\n",
      "# 'Courtesy Report':\n",
      "  . Most correlated unigrams:\n",
      ". report\n",
      ". courtesy\n",
      "  . Most correlated bigrams:\n",
      ". theft locked\n",
      ". courtesy report\n",
      "# 'Disorderly Conduct':\n",
      "  . Most correlated unigrams:\n",
      ". threats\n",
      ". terrorist\n",
      "  . Most correlated bigrams:\n",
      ". influence public\n",
      ". terrorist threats\n",
      "# 'Drug Offense':\n",
      "  . Most correlated unigrams:\n",
      ". possession\n",
      ". sale\n",
      "  . Most correlated bigrams:\n",
      ". paraphernalia possession\n",
      ". possession sale\n",
      "# 'Embezzlement':\n",
      "  . Most correlated unigrams:\n",
      ". employee\n",
      ". embezzled\n",
      "  . Most correlated bigrams:\n",
      ". embezzlement grand\n",
      ". grand theft\n",
      "# 'Family Offense':\n",
      "  . Most correlated unigrams:\n",
      ". mental\n",
      ". physical\n",
      "  . Most correlated bigrams:\n",
      ". suffering death\n",
      ". physical pain\n",
      "# 'Fire Report':\n",
      "  . Most correlated unigrams:\n",
      ". theft\n",
      ". report\n",
      "  . Most correlated bigrams:\n",
      ". locked vehicle\n",
      ". theft locked\n",
      "# 'Forgery And Counterfeiting':\n",
      "  . Most correlated unigrams:\n",
      ". checks\n",
      ". forgery\n",
      "  . Most correlated bigrams:\n",
      ". false evidence\n",
      ". registration false\n",
      "# 'Fraud':\n",
      "  . Most correlated unigrams:\n",
      ". personation\n",
      ". fraudulent\n",
      "  . Most correlated bigrams:\n",
      ". money property\n",
      ". false personation\n",
      "# 'Larceny Theft':\n",
      "  . Most correlated unigrams:\n",
      ". 950\n",
      ". theft\n",
      "  . Most correlated bigrams:\n",
      ". theft locked\n",
      ". locked vehicle\n",
      "# 'Lost Property':\n",
      "  . Most correlated unigrams:\n",
      ". property\n",
      ". lost\n",
      "  . Most correlated bigrams:\n",
      ". theft locked\n",
      ". lost property\n",
      "# 'Missing Person':\n",
      "  . Most correlated unigrams:\n",
      ". missing\n",
      ". person\n",
      "  . Most correlated bigrams:\n",
      ". theft locked\n",
      ". missing adult\n",
      "# 'Motor Vehicle Theft':\n",
      "  . Most correlated unigrams:\n",
      ". auto\n",
      ". stolen\n",
      "  . Most correlated bigrams:\n",
      ". stolen auto\n",
      ". vehicle stolen\n",
      "# 'Offences Against The Family And Children':\n",
      "  . Most correlated unigrams:\n",
      ". restraining\n",
      ". order\n",
      "  . Most correlated bigrams:\n",
      ". violence secondary\n",
      ". restraining order\n",
      "# 'Prostitution':\n",
      "  . Most correlated unigrams:\n",
      ". purpose\n",
      ". prostitution\n",
      "  . Most correlated bigrams:\n",
      ". purpose prostitution\n",
      ". loitering purpose\n",
      "# 'Recovered Vehicle':\n",
      "  . Most correlated unigrams:\n",
      ". auto\n",
      ". recovered\n",
      "  . Most correlated bigrams:\n",
      ". recovered auto\n",
      ". vehicle recovered\n",
      "# 'Robbery':\n",
      "  . Most correlated unigrams:\n",
      ". force\n",
      ". robbery\n",
      "  . Most correlated bigrams:\n",
      ". street public\n",
      ". robbery force\n",
      "# 'Sex Offense':\n",
      "  . Most correlated unigrams:\n",
      ". abuse\n",
      ". pornography\n",
      "  . Most correlated bigrams:\n",
      ". abuse pornography\n",
      ". child abuse\n",
      "# 'Stolen Property':\n",
      "  . Most correlated unigrams:\n",
      ". receiving\n",
      ". knowledge\n",
      "  . Most correlated bigrams:\n",
      ". property possession\n",
      ". knowledge receiving\n",
      "# 'Suspicious Occ':\n",
      "  . Most correlated unigrams:\n",
      ". occurrence\n",
      ". suspicious\n",
      "  . Most correlated bigrams:\n",
      ". suspicious act\n",
      ". suspicious occurrence\n",
      "# 'Traffic Collision':\n",
      "  . Most correlated unigrams:\n",
      ". hit\n",
      ". collision\n",
      "  . Most correlated bigrams:\n",
      ". collision hit\n",
      ". traffic collision\n",
      "# 'Traffic Violation Arrest':\n",
      "  . Most correlated unigrams:\n",
      ". violation\n",
      ". traffic\n",
      "  . Most correlated bigrams:\n",
      ". traffic violation\n",
      ". violation arrest\n",
      "# 'Vandalism':\n",
      "  . Most correlated unigrams:\n",
      ". vehicle\n",
      ". tampering\n",
      "  . Most correlated bigrams:\n",
      ". locked vehicle\n",
      ". vehicle tampering\n",
      "# 'Vehicle Impounded':\n",
      "  . Most correlated unigrams:\n",
      ". vehicle\n",
      ". impounded\n",
      "  . Most correlated bigrams:\n",
      ". theft locked\n",
      ". vehicle impounded\n",
      "# 'Warrant':\n",
      "  . Most correlated unigrams:\n",
      ". local\n",
      ". warrant\n",
      "  . Most correlated bigrams:\n",
      ". arrest local\n",
      ". warrant arrest\n",
      "# 'Weapons Carrying Etc':\n",
      "  . Most correlated unigrams:\n",
      ". firearm\n",
      ". prohibited\n",
      "  . Most correlated bigrams:\n",
      ". prohibited person\n",
      ". firearm possession\n",
      "# 'Weapons Offense':\n",
      "  . Most correlated unigrams:\n",
      ". imitation\n",
      ". deadly\n",
      "  . Most correlated bigrams:\n",
      ". deadly imitation\n",
      ". weapon deadly\n"
     ]
    }
   ],
   "source": [
    "# Use sklearn.feature_selection.chi2 to find the terms that are the most correlated with each of the Category:\n",
    "from sklearn.feature_selection import chi2\n",
    "N = 2\n",
    "for Category, category_id in sorted(category_to_id.items()):\n",
    "  features_chi2 = chi2(features, labels == category_id)\n",
    "  indices = np.argsort(features_chi2[0])\n",
    "  feature_names = np.array(tfidf.get_feature_names())[indices]\n",
    "  unigrams = [v for v in feature_names if len(v.split(' ')) == 1]\n",
    "  bigrams = [v for v in feature_names if len(v.split(' ')) == 2]\n",
    "  print(\"# '{}':\".format(Category))\n",
    "  print(\"  . Most correlated unigrams:\\n. {}\".format('\\n. '.join(unigrams[-N:])))\n",
    "  print(\"  . Most correlated bigrams:\\n. {}\".format('\\n. '.join(bigrams[-N:])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Use Naive Bayes Classifier to find word counts is the multinomial variant\n",
    "# https://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.MultinomialNB.html\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.feature_extraction.text import TfidfTransformer\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "X_train, X_test, y_train, y_test = train_test_split(df1['Description'], df1['Category'], random_state = 0)\n",
    "count_vect = CountVectorizer()\n",
    "X_train_counts = count_vect.fit_transform(X_train)\n",
    "tfidf_transformer = TfidfTransformer()\n",
    "X_train_tfidf = tfidf_transformer.fit_transform(X_train_counts)\n",
    "clf = MultinomialNB().fit(X_train_tfidf, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Category                      Larceny Theft\n",
       "Description    Theft, Other Property, >$950\n",
       "category_id                               4\n",
       "Name: 1000, dtype: object"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Take 1 sample record in row 1000\n",
    "df1.loc[1000,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Larceny Theft']\n"
     ]
    }
   ],
   "source": [
    "# Use the Description text to try to predict Category\n",
    "print(clf.predict(count_vect.transform([\"Theft, Other Property, >$950\"])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:469: FutureWarning: Default multi_class will be changed to 'auto' in 0.22. Specify the multi_class option to silence this warning.\n",
      "  \"this warning.\", FutureWarning)\n",
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:469: FutureWarning: Default multi_class will be changed to 'auto' in 0.22. Specify the multi_class option to silence this warning.\n",
      "  \"this warning.\", FutureWarning)\n",
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:469: FutureWarning: Default multi_class will be changed to 'auto' in 0.22. Specify the multi_class option to silence this warning.\n",
      "  \"this warning.\", FutureWarning)\n",
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:469: FutureWarning: Default multi_class will be changed to 'auto' in 0.22. Specify the multi_class option to silence this warning.\n",
      "  \"this warning.\", FutureWarning)\n",
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
      "  FutureWarning)\n",
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:469: FutureWarning: Default multi_class will be changed to 'auto' in 0.22. Specify the multi_class option to silence this warning.\n",
      "  \"this warning.\", FutureWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Try experiment with different machine learning models, evaluate their accuracy and find the source of any potential issues.\n",
    "# We will benchmark the following four models/algorithms, Random Forest, Linear Support Vector Machine, (Multinomial) Naive Bayes, Logistic Regression.\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.model_selection import cross_val_score\n",
    "models = [\n",
    "    RandomForestClassifier(n_estimators=200, max_depth=3, random_state=0),\n",
    "    LinearSVC(),\n",
    "    MultinomialNB(),\n",
    "    LogisticRegression(random_state=0),\n",
    "]\n",
    "CV = 5\n",
    "cv_df = pd.DataFrame(index=range(CV * len(models)))\n",
    "entries = []\n",
    "for model in models:\n",
    "  model_name = model.__class__.__name__\n",
    "  accuracies = cross_val_score(model, features, labels, scoring='accuracy', cv=CV)\n",
    "  for fold_idx, accuracy in enumerate(accuracies):\n",
    "    entries.append((model_name, fold_idx, accuracy))\n",
    "cv_df = pd.DataFrame(entries, columns=['model_name', 'fold_idx', 'accuracy'])\n",
    "import seaborn as sns\n",
    "sns.boxplot(x='model_name', y='accuracy', data=cv_df)\n",
    "sns.stripplot(x='model_name', y='accuracy', data=cv_df, \n",
    "              size=8, jitter=True, edgecolor=\"gray\", linewidth=2)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "model_name\n",
       "LinearSVC                 0.999904\n",
       "LogisticRegression        0.999785\n",
       "MultinomialNB             0.997088\n",
       "RandomForestClassifier    0.477412\n",
       "Name: accuracy, dtype: float64"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Show each of models/algorithm accuracy rate\n",
    "cv_df.groupby('model_name').accuracy.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Show the confusion matrix for model evaluation\n",
    "#model = LinearSVC()\n",
    "model = RandomForestClassifier(n_estimators=200, max_depth=3, random_state=0)\n",
    "X_train, X_test, y_train, y_test, indices_train, indices_test = train_test_split(features, labels, df1.index, test_size=0.33, random_state=0)\n",
    "model.fit(X_train, y_train)\n",
    "y_pred = model.predict(X_test)\n",
    "from sklearn.metrics import confusion_matrix\n",
    "conf_mat = confusion_matrix(y_test, y_pred)\n",
    "fig, ax = plt.subplots(figsize=(10,10))\n",
    "sns.heatmap(conf_mat, annot=True, fmt='d',\n",
    "            xticklabels=category_id_df1.Category.values, yticklabels=category_id_df1.Category.values)\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "'Larceny Theft' predicted as 'Offences Against The Family And Children' : 42 examples.\n"
     ]
    },
    {
     "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>Category</th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>15321</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57107</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>148342</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>111466</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13838</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Looting during State of Emergency</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>154407</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32040</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>145080</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>129617</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>124802</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>148068</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98571</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39490</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>118906</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32082</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>158372</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125837</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47114</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99924</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>152167</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108895</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132720</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>158079</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>190275</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>167837</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>128212</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11487</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>188461</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43872</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33994</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>190474</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82047</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74160</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>174584</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>160044</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56322</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139078</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108087</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>171950</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86355</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>154691</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7580</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Embezzlement from Dependent or Elder Adult by ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Category                                        Description\n",
       "15321   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "57107   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "148342  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "111466  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "13838   Larceny Theft                  Looting during State of Emergency\n",
       "154407  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "32040   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "145080  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "129617  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "124802  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "148068  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "98571   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "39490   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "118906  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "32082   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "158372  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "125837  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "47114   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "99924   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "152167  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "108895  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "132720  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "158079  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "190275  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "167837  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "128212  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "11487   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "188461  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "43872   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "33994   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "190474  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "82047   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "74160   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "174584  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "160044  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "56322   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "139078  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "108087  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "171950  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "86355   Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "154691  Larceny Theft  Embezzlement from Dependent or Elder Adult by ...\n",
       "7580    Larceny Theft  Embezzlement from Dependent or Elder Adult by ..."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "'Sex Offense' predicted as 'Disorderly Conduct' : 7 examples.\n"
     ]
    },
    {
     "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>Category</th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>169232</th>\n",
       "      <td>Sex Offense</td>\n",
       "      <td>Phone Calls, Obscene</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>143302</th>\n",
       "      <td>Sex Offense</td>\n",
       "      <td>Phone Calls, Obscene</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30630</th>\n",
       "      <td>Sex Offense</td>\n",
       "      <td>Phone Calls, Obscene</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>175235</th>\n",
       "      <td>Sex Offense</td>\n",
       "      <td>Phone Calls, Obscene</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207412</th>\n",
       "      <td>Sex Offense</td>\n",
       "      <td>Phone Calls, Obscene</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40093</th>\n",
       "      <td>Sex Offense</td>\n",
       "      <td>Phone Calls, Obscene</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81631</th>\n",
       "      <td>Sex Offense</td>\n",
       "      <td>Phone Calls, Obscene</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Category           Description\n",
       "169232  Sex Offense  Phone Calls, Obscene\n",
       "143302  Sex Offense  Phone Calls, Obscene\n",
       "30630   Sex Offense  Phone Calls, Obscene\n",
       "175235  Sex Offense  Phone Calls, Obscene\n",
       "207412  Sex Offense  Phone Calls, Obscene\n",
       "40093   Sex Offense  Phone Calls, Obscene\n",
       "81631   Sex Offense  Phone Calls, Obscene"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "'Fire Report' predicted as 'Courtesy Report' : 78 examples.\n"
     ]
    },
    {
     "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>Category</th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>190882</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>179716</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>149899</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10796</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38154</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61502</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96171</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>122269</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>183202</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20480</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13959</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97195</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48143</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>213333</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>151258</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>182293</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>111885</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>167312</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>212701</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>113437</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>64674</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92040</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>151462</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108629</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>198494</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>113178</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47781</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63343</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>216248</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161197</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>110056</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>85492</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36859</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>198853</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89760</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91729</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>145562</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>167625</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92490</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>145686</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>186008</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>183198</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>179890</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9497</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>134238</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>163570</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>110317</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>189152</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>160467</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207320</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15952</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>195316</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>124334</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7025</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>210993</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>190167</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47371</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>172116</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>177839</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15690</th>\n",
       "      <td>Fire Report</td>\n",
       "      <td>Fire Report</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>78 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           Category  Description\n",
       "190882  Fire Report  Fire Report\n",
       "179716  Fire Report  Fire Report\n",
       "149899  Fire Report  Fire Report\n",
       "10796   Fire Report  Fire Report\n",
       "38154   Fire Report  Fire Report\n",
       "61502   Fire Report  Fire Report\n",
       "96171   Fire Report  Fire Report\n",
       "122269  Fire Report  Fire Report\n",
       "183202  Fire Report  Fire Report\n",
       "20480   Fire Report  Fire Report\n",
       "13959   Fire Report  Fire Report\n",
       "97195   Fire Report  Fire Report\n",
       "48143   Fire Report  Fire Report\n",
       "213333  Fire Report  Fire Report\n",
       "151258  Fire Report  Fire Report\n",
       "182293  Fire Report  Fire Report\n",
       "111885  Fire Report  Fire Report\n",
       "167312  Fire Report  Fire Report\n",
       "212701  Fire Report  Fire Report\n",
       "113437  Fire Report  Fire Report\n",
       "64674   Fire Report  Fire Report\n",
       "92040   Fire Report  Fire Report\n",
       "151462  Fire Report  Fire Report\n",
       "108629  Fire Report  Fire Report\n",
       "198494  Fire Report  Fire Report\n",
       "113178  Fire Report  Fire Report\n",
       "47781   Fire Report  Fire Report\n",
       "63343   Fire Report  Fire Report\n",
       "216248  Fire Report  Fire Report\n",
       "161197  Fire Report  Fire Report\n",
       "...             ...          ...\n",
       "110056  Fire Report  Fire Report\n",
       "85492   Fire Report  Fire Report\n",
       "36859   Fire Report  Fire Report\n",
       "198853  Fire Report  Fire Report\n",
       "89760   Fire Report  Fire Report\n",
       "91729   Fire Report  Fire Report\n",
       "145562  Fire Report  Fire Report\n",
       "167625  Fire Report  Fire Report\n",
       "92490   Fire Report  Fire Report\n",
       "145686  Fire Report  Fire Report\n",
       "186008  Fire Report  Fire Report\n",
       "183198  Fire Report  Fire Report\n",
       "179890  Fire Report  Fire Report\n",
       "9497    Fire Report  Fire Report\n",
       "134238  Fire Report  Fire Report\n",
       "163570  Fire Report  Fire Report\n",
       "110317  Fire Report  Fire Report\n",
       "189152  Fire Report  Fire Report\n",
       "160467  Fire Report  Fire Report\n",
       "207320  Fire Report  Fire Report\n",
       "15952   Fire Report  Fire Report\n",
       "195316  Fire Report  Fire Report\n",
       "124334  Fire Report  Fire Report\n",
       "7025    Fire Report  Fire Report\n",
       "210993  Fire Report  Fire Report\n",
       "190167  Fire Report  Fire Report\n",
       "47371   Fire Report  Fire Report\n",
       "172116  Fire Report  Fire Report\n",
       "177839  Fire Report  Fire Report\n",
       "15690   Fire Report  Fire Report\n",
       "\n",
       "[78 rows x 2 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "'Disorderly Conduct' predicted as 'Prostitution' : 14 examples.\n"
     ]
    },
    {
     "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>Category</th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>54212</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Engaging In Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>213561</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Loitering Around Public Toilet For Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>204261</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Engaging In Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>110717</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Engaging In Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132433</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Loitering Around Public Toilet For Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181552</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Engaging In Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>157723</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Loitering Around Public Toilet For Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78568</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Engaging In Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>183050</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Engaging In Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>191014</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Engaging In Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>106549</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Engaging In Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60579</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Engaging In Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102818</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Engaging In Lewd Act</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73725</th>\n",
       "      <td>Disorderly Conduct</td>\n",
       "      <td>Engaging In Lewd Act</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  Category                                  Description\n",
       "54212   Disorderly Conduct                         Engaging In Lewd Act\n",
       "213561  Disorderly Conduct  Loitering Around Public Toilet For Lewd Act\n",
       "204261  Disorderly Conduct                         Engaging In Lewd Act\n",
       "110717  Disorderly Conduct                         Engaging In Lewd Act\n",
       "132433  Disorderly Conduct  Loitering Around Public Toilet For Lewd Act\n",
       "181552  Disorderly Conduct                         Engaging In Lewd Act\n",
       "157723  Disorderly Conduct  Loitering Around Public Toilet For Lewd Act\n",
       "78568   Disorderly Conduct                         Engaging In Lewd Act\n",
       "183050  Disorderly Conduct                         Engaging In Lewd Act\n",
       "191014  Disorderly Conduct                         Engaging In Lewd Act\n",
       "106549  Disorderly Conduct                         Engaging In Lewd Act\n",
       "60579   Disorderly Conduct                         Engaging In Lewd Act\n",
       "102818  Disorderly Conduct                         Engaging In Lewd Act\n",
       "73725   Disorderly Conduct                         Engaging In Lewd Act"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "'Larceny Theft' predicted as 'Fire Report' : 10 examples.\n"
     ]
    },
    {
     "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>Category</th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>172099</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft of Telecommunication Services, incl. Clo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150184</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft, Coin Operated Machine, &lt;$50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131517</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft, Coin Operated Machine, &gt;$950</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150286</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft, Coin Operated Machine, $200-$950</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>135250</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft, Coin Operated Machine, &lt;$50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69019</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft, Coin Operated Machine, $200-$950</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>134514</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft, Coin Operated Machine, $200-$950</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>188435</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft of Telecommunication Services, incl. Clo...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>217448</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft, Coin Operated Machine, &lt;$50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108933</th>\n",
       "      <td>Larceny Theft</td>\n",
       "      <td>Theft of Telecommunication Services, incl. Clo...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Category                                        Description\n",
       "172099  Larceny Theft  Theft of Telecommunication Services, incl. Clo...\n",
       "150184  Larceny Theft                 Theft, Coin Operated Machine, <$50\n",
       "131517  Larceny Theft                Theft, Coin Operated Machine, >$950\n",
       "150286  Larceny Theft            Theft, Coin Operated Machine, $200-$950\n",
       "135250  Larceny Theft                 Theft, Coin Operated Machine, <$50\n",
       "69019   Larceny Theft            Theft, Coin Operated Machine, $200-$950\n",
       "134514  Larceny Theft            Theft, Coin Operated Machine, $200-$950\n",
       "188435  Larceny Theft  Theft of Telecommunication Services, incl. Clo...\n",
       "217448  Larceny Theft                 Theft, Coin Operated Machine, <$50\n",
       "108933  Larceny Theft  Theft of Telecommunication Services, incl. Clo..."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "'Fraud' predicted as 'Fire Report' : 7 examples.\n"
     ]
    },
    {
     "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>Category</th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>59607</th>\n",
       "      <td>Fraud</td>\n",
       "      <td>Telephone or Telegraph Message, Sending False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>745</th>\n",
       "      <td>Fraud</td>\n",
       "      <td>Telephone or Telegraph Message, Sending False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>113133</th>\n",
       "      <td>Fraud</td>\n",
       "      <td>Telephone or Telegraph Message, Sending False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126392</th>\n",
       "      <td>Fraud</td>\n",
       "      <td>Audiovisual (Video or Sound) Recordings, Unaut...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68923</th>\n",
       "      <td>Fraud</td>\n",
       "      <td>Audiovisual (Video or Sound) Recordings, Unaut...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93316</th>\n",
       "      <td>Fraud</td>\n",
       "      <td>Audiovisual (Video or Sound) Recordings, Unaut...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>205557</th>\n",
       "      <td>Fraud</td>\n",
       "      <td>Audiovisual (Video or Sound) Recordings, Unaut...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Category                                        Description\n",
       "59607     Fraud      Telephone or Telegraph Message, Sending False\n",
       "745       Fraud      Telephone or Telegraph Message, Sending False\n",
       "113133    Fraud      Telephone or Telegraph Message, Sending False\n",
       "126392    Fraud  Audiovisual (Video or Sound) Recordings, Unaut...\n",
       "68923     Fraud  Audiovisual (Video or Sound) Recordings, Unaut...\n",
       "93316     Fraud  Audiovisual (Video or Sound) Recordings, Unaut...\n",
       "205557    Fraud  Audiovisual (Video or Sound) Recordings, Unaut..."
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# Try to use test data to perform prediction\n",
    "from IPython.display import display\n",
    "\n",
    "for predicted in category_id_df1.category_id:\n",
    "  for actual in category_id_df1.category_id:\n",
    "    if predicted != actual and conf_mat[actual, predicted] >= 6:\n",
    "      print(\"'{}' predicted as '{}' : {} examples.\".format(id_to_category[actual], id_to_category[predicted], conf_mat[actual, predicted]))\n",
    "      display(df1.loc[indices_test[(y_test == actual) & (y_pred == predicted)]][['Category', 'Description']])\n",
    "      print('')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                          precision    recall  f1-score   support\n",
      "\n",
      "                          Missing Person       0.00      0.00      0.00      2288\n",
      "                         Stolen Property       0.00      0.00      0.00       502\n",
      "Offences Against The Family And Children       0.00      0.00      0.00      1938\n",
      "                           Lost Property       0.00      0.00      0.00      2915\n",
      "                           Larceny Theft       0.43      1.00      0.60     28887\n",
      "                                 Warrant       1.00      0.88      0.94      3588\n",
      "                                   Fraud       0.00      0.00      0.00      2934\n",
      "                                Burglary       1.00      0.52      0.69      4257\n",
      "                Traffic Violation Arrest       0.00      0.00      0.00      1258\n",
      "                    Weapons Carrying Etc       0.00      0.00      0.00       523\n",
      "                                 Assault       0.00      0.00      0.00      5662\n",
      "                            Drug Offense       0.00      0.00      0.00      2533\n",
      "                     Motor Vehicle Theft       0.00      0.00      0.00      3550\n",
      "                       Recovered Vehicle       0.00      0.00      0.00      2182\n",
      "                                 Robbery       0.00      0.00      0.00      2375\n",
      "                          Suspicious Occ       0.00      0.00      0.00      1847\n",
      "                      Disorderly Conduct       0.00      0.00      0.00      1797\n",
      "                         Weapons Offense       0.00      0.00      0.00       526\n",
      "                               Vandalism       0.00      0.00      0.00       122\n",
      "                            Embezzlement       0.00      0.00      0.00       151\n",
      "                         Courtesy Report       0.00      0.00      0.00       283\n",
      "                       Traffic Collision       0.00      0.00      0.00       197\n",
      "                            Prostitution       0.00      0.00      0.00       249\n",
      "              Forgery And Counterfeiting       0.00      0.00      0.00       424\n",
      "                             Sex Offense       0.00      0.00      0.00       307\n",
      "                                   Arson       0.00      0.00      0.00       237\n",
      "                       Vehicle Impounded       0.00      0.00      0.00        78\n",
      "                         Civil Sidewalks       0.00      0.00      0.00       261\n",
      "                             Fire Report       0.00      0.00      0.00        78\n",
      "                          Family Offense       0.00      0.00      0.00       136\n",
      "\n",
      "                                accuracy                           0.48     72085\n",
      "                               macro avg       0.08      0.08      0.07     72085\n",
      "                            weighted avg       0.28      0.48      0.33     72085\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\easlai\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\sklearn\\metrics\\classification.py:1437: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    }
   ],
   "source": [
    "# Print out classification report for each class/category\n",
    "from sklearn import metrics\n",
    "print(metrics.classification_report(y_test, y_pred, target_names=df['Category'].unique()))"
   ]
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}