NBExample.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import the required libraries\n",
    "import os\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import matplotlib.pyplot as plt  # for creating plots of the data\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import confusion_matrix, roc_curve, auc, roc_auc_score\n",
    "import itertools\n",
    "\n",
    "# Import our own library, which has some utility functions and the Naive Bayes Classifier we created\n",
    "import CatNB \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The dataset is in the Data folder (this is the same data as for the Logistic Regression Example)\n",
    "DATASET_PATH = 'Data/'\n",
    "\n",
    "# Read the data from the CSV file\n",
    "data_path = os.path.join(DATASET_PATH, 'census1.csv')\n",
    "censusData = pd.read_csv(data_path, na_values = [\"?\", \" ?\"], encoding = \"utf-7\") # we know that libreOffice is using utf-7 encoding\n",
    "\n",
    "# Because the CSV doesn't contain any header, we add column names \n",
    "# using the description from the original dataset website \n",
    "# http://mlr.cs.umass.edu/ml/machine-learning-databases/adult\n",
    "\n",
    "censusData.columns = [\n",
    "    \"Age\", \"WorkClass\", \"FnlWgt\",\n",
    "    \"Education\", \"EducationNum\", \"MaritalStatus\",\n",
    "    \"Occupation\", \"Relationship\", \"Race\", \"Gender\", \n",
    "    \"CapitalGain\", \"CapitalLoss\", \"HoursPerWeek\", \"NativeCountry\", \"Income\"]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   Age          WorkClass  FnlWgt   Education  EducationNum  \\\n",
      "0   50   Self-emp-not-inc   83311   Bachelors            13   \n",
      "1   38            Private  215646     HS-grad             9   \n",
      "2   53            Private  234721        11th             7   \n",
      "3   28            Private  338409   Bachelors            13   \n",
      "4   37            Private  284582     Masters            14   \n",
      "\n",
      "         MaritalStatus          Occupation    Relationship    Race   Gender  \\\n",
      "0   Married-civ-spouse     Exec-managerial         Husband   White     Male   \n",
      "1             Divorced   Handlers-cleaners   Not-in-family   White     Male   \n",
      "2   Married-civ-spouse   Handlers-cleaners         Husband   Black     Male   \n",
      "3   Married-civ-spouse      Prof-specialty            Wife   Black   Female   \n",
      "4   Married-civ-spouse     Exec-managerial            Wife   White   Female   \n",
      "\n",
      "   CapitalGain  CapitalLoss  HoursPerWeek   NativeCountry  Income  \n",
      "0            0            0            13   United-States   <=50K  \n",
      "1            0            0            40   United-States   <=50K  \n",
      "2            0            0            40   United-States   <=50K  \n",
      "3            0            0            40            Cuba   <=50K  \n",
      "4            0            0            40   United-States   <=50K  \n",
      "                Age        FnlWgt  EducationNum   CapitalGain   CapitalLoss  \\\n",
      "count  32560.000000  3.256000e+04  32560.000000  32560.000000  32560.000000   \n",
      "mean      38.581634  1.897818e+05     10.080590   1077.615172     87.306511   \n",
      "std       13.640642  1.055498e+05      2.572709   7385.402999    402.966116   \n",
      "min       17.000000  1.228500e+04      1.000000      0.000000      0.000000   \n",
      "25%       28.000000  1.178315e+05      9.000000      0.000000      0.000000   \n",
      "50%       37.000000  1.783630e+05     10.000000      0.000000      0.000000   \n",
      "75%       48.000000  2.370545e+05     12.000000      0.000000      0.000000   \n",
      "max       90.000000  1.484705e+06     16.000000  99999.000000   4356.000000   \n",
      "\n",
      "       HoursPerWeek  \n",
      "count  32560.000000  \n",
      "mean      40.437469  \n",
      "std       12.347618  \n",
      "min        1.000000  \n",
      "25%       40.000000  \n",
      "50%       40.000000  \n",
      "75%       45.000000  \n",
      "max       99.000000  \n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 32560 entries, 0 to 32559\n",
      "Data columns (total 15 columns):\n",
      "Age              32560 non-null int64\n",
      "WorkClass        30724 non-null object\n",
      "FnlWgt           32560 non-null int64\n",
      "Education        32560 non-null object\n",
      "EducationNum     32560 non-null int64\n",
      "MaritalStatus    32560 non-null object\n",
      "Occupation       30717 non-null object\n",
      "Relationship     32560 non-null object\n",
      "Race             32560 non-null object\n",
      "Gender           32560 non-null object\n",
      "CapitalGain      32560 non-null int64\n",
      "CapitalLoss      32560 non-null int64\n",
      "HoursPerWeek     32560 non-null int64\n",
      "NativeCountry    31977 non-null object\n",
      "Income           32560 non-null object\n",
      "dtypes: int64(6), object(9)\n",
      "memory usage: 3.7+ MB\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "# Get to know the data with head, describe, and info methods\n",
    "\n",
    "# Visualise a table with the first rows of the dataset, to\n",
    "# better understand the data format\n",
    "print(censusData.head())\n",
    "print(censusData.describe())\n",
    "print(censusData.info())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = censusData.drop([\"Income\"], axis = 1)\n",
    "Y = censusData.Income\n",
    "# Now, lets split the data into train and test sets:\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size = 0.2, random_state = 18)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "CNB = CatNB.CategoricalNaiveBayes()  # instantiate a Categorical Naive Bayes object "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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>WorkClass</th>\n",
       "      <th>Education</th>\n",
       "      <th>MaritalStatus</th>\n",
       "      <th>Occupation</th>\n",
       "      <th>Relationship</th>\n",
       "      <th>Race</th>\n",
       "      <th>Gender</th>\n",
       "      <th>NativeCountry</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Income</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>&lt;=50K</th>\n",
       "      <td>18429</td>\n",
       "      <td>19740</td>\n",
       "      <td>19740</td>\n",
       "      <td>18422</td>\n",
       "      <td>19740</td>\n",
       "      <td>19740</td>\n",
       "      <td>19740</td>\n",
       "      <td>19401</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>&gt;50K</th>\n",
       "      <td>6158</td>\n",
       "      <td>6308</td>\n",
       "      <td>6308</td>\n",
       "      <td>6158</td>\n",
       "      <td>6308</td>\n",
       "      <td>6308</td>\n",
       "      <td>6308</td>\n",
       "      <td>6196</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        WorkClass  Education  MaritalStatus  Occupation  Relationship   Race  \\\n",
       "Income                                                                         \n",
       "<=50K       18429      19740          19740       18422         19740  19740   \n",
       ">50K         6158       6308           6308        6158          6308   6308   \n",
       "\n",
       "        Gender  NativeCountry  \n",
       "Income                         \n",
       "<=50K    19740          19401  \n",
       ">50K      6308           6196  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainData = X_train.assign(Income = Y_train)\n",
    "model = CNB.fit(trainData)\n",
    "model.aggregateTotals"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Das Modell hat sich aus den Zensus Daten die Kategorischen Attributen ausgesucht.  Wir können hier schon eine Vorhersage machen, und dann eine Confusion Matrix mit dem Resultat erstellen."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "testData = X_test.assign(Income = Y_test)\n",
    "Y_pred_NB = model.predict(testData)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n",
      "[[4078  901]\n",
      " [ 382 1151]]\n",
      "Normalized confusion matrix\n",
      "[[0.82 0.18]\n",
      " [0.25 0.75]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "trueY = Y_test == \" >50K\"\n",
    "predY = Y_pred_NB.MaxClass == \">50K\"\n",
    "\n",
    "\n",
    "# Compute confusion matrix\n",
    "cnf_matrix = confusion_matrix(trueY, predY)\n",
    "\n",
    "np.set_printoptions(precision=2)\n",
    "\n",
    "# Plot non-normalized confusion matrix\n",
    "plt.figure()\n",
    "CatNB.plot_confusion_matrix(cnf_matrix, classes=[\"<=50K\", \">50K\"],\n",
    "                      title='Confusion matrix, without normalization')\n",
    "\n",
    "# Plot normalized confusion matrix\n",
    "plt.figure()\n",
    "CatNB.plot_confusion_matrix(cnf_matrix, classes=[\"<=50K\", \">50K\"], normalize = True,\n",
    "                      title='Confusion matrix, with normalization')"
   ]
  }
 ],
 "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}