gaussian_nb.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package punkt to\n",
      "[nltk_data]     C:\\Users\\jande\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n",
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     C:\\Users\\jande\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "######## Import of necessary packages ########\n",
    "# CountVectorizer for creating the BoW matrix\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "\n",
    "# Cross validation for model evaluation\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.model_selection import cross_val_score, train_test_split\n",
    "\n",
    "# Pandas for dataset preparation\n",
    "import pandas as pd\n",
    "\n",
    "# GaussianNB as Naive Bayes implementation for model\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "\n",
    "import re\n",
    "\n",
    "# Import stopwords word_tokenize from Natural Language Toolkit for cleaning the text columns of the dataset and splitting sentences into single words\n",
    "import nltk\n",
    "from nltk.corpus import stopwords\n",
    "##############################################\n",
    "\n",
    "\n",
    "# Definition of column names of the dataset\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "columnname_account = 'Auftragskonto'\n",
    "columnname_orderdate = 'Buchungstag'\n",
    "columnname_valutadate = 'Valutadatum'\n",
    "columnname_usage = 'Verwendungszweck'\n",
    "columnname_text = 'Buchungstext'\n",
    "columnname_target_sender = 'Beguenstigter/Zahlungspflichtiger'\n",
    "columnname_depotnumber = 'Kontonummer'\n",
    "columnname_blz = 'BLZ'\n",
    "columnname_amount = 'Betrag'\n",
    "columnname_currency = 'Waehrung'\n",
    "columnname_label = 'label'\n",
    "\n",
    "# Definition of the the column names for the training dataframe\n",
    "columnname_final_messages = 'messages'\n",
    "columnname_final_class = 'class'\n",
    "\n",
    "dataset_file = \"dataset.csv\"\n",
    "dataset_separator = \";\"\n",
    "\n",
    "# Percentage of the training set size based on the dataset size\n",
    "training_set_size = 0.8\n",
    "\n",
    "nltk.download('punkt')\n",
    "nltk.download('stopwords')\n",
    "\n",
    "# Set displayed rows\n",
    "pd.set_option('display.max_rows', 10)\n",
    "\n",
    "general_bag_of_words = list()"
   ]
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Data Load"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Complete dataset size: 209\n"
     ]
    }
   ],
   "source": [
    "data = pd.read_csv(dataset_file, sep=dataset_separator)\n",
    "print(\"Complete dataset size: \" +str(len(data)))"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Raw Data Shape"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "####################### Raw Data #######################\n",
      "                     0  Null Values  count          mean        std  \\\n",
      "Unnamed: 0       int64            0  209.0  1.040000e+02  60.477268   \n",
      "Auftragskonto  float64           41  168.0  8.999021e+07   4.077940   \n",
      "Buchungstag     object            0    NaN           NaN        NaN   \n",
      "Valutadatum     object            0    NaN           NaN        NaN   \n",
      "Buchungstext    object            0    NaN           NaN        NaN   \n",
      "...                ...          ...    ...           ...        ...   \n",
      "Kontonummer     object            1    NaN           NaN        NaN   \n",
      "BLZ             object            1    NaN           NaN        NaN   \n",
      "Betrag          object            0    NaN           NaN        NaN   \n",
      "Waehrung        object            0    NaN           NaN        NaN   \n",
      "label           object            0    NaN           NaN        NaN   \n",
      "\n",
      "                      min         25%         50%         75%         max  \n",
      "Unnamed: 0            0.0        52.0       104.0       156.0       208.0  \n",
      "Auftragskonto  89990201.0  89990201.0  89990210.0  89990210.0  89990210.0  \n",
      "Buchungstag           NaN         NaN         NaN         NaN         NaN  \n",
      "Valutadatum           NaN         NaN         NaN         NaN         NaN  \n",
      "Buchungstext          NaN         NaN         NaN         NaN         NaN  \n",
      "...                   ...         ...         ...         ...         ...  \n",
      "Kontonummer           NaN         NaN         NaN         NaN         NaN  \n",
      "BLZ                   NaN         NaN         NaN         NaN         NaN  \n",
      "Betrag                NaN         NaN         NaN         NaN         NaN  \n",
      "Waehrung              NaN         NaN         NaN         NaN         NaN  \n",
      "label                 NaN         NaN         NaN         NaN         NaN  \n",
      "\n",
      "[12 rows x 10 columns]\n"
     ]
    }
   ],
   "source": [
    "print(\"####################### Raw Data #######################\")\n",
    "data_raw_summary = pd.DataFrame(data.dtypes)\n",
    "data_raw_statistics = data.describe().transpose()\n",
    "data_raw_summary[\"Null Values\"] = data.isnull().sum()\n",
    "data_raw_summary = pd.concat([data_raw_summary,data_raw_statistics], axis=1)\n",
    "print(data_raw_summary)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Data Preparation"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting preparation of data...\n",
      "Preparing data...\n",
      "Preparing text featues...\n",
      "Prepared data!\n"
     ]
    }
   ],
   "source": [
    "def removePunctuationAndSpecialCharacters(text):\n",
    "    return text.replace('.', '').replace('&', '').replace(':', '').replace('/', ' ').replace('-',' ')\n",
    "\n",
    "def removeSpecialWords(text):\n",
    "    # Remove End-to-End-Ref\n",
    "    text = text.replace('end-to-end-ref', ' ')\n",
    "    # Remove NotProvided\n",
    "    text = text.replace('notprovided ', ' ')\n",
    "    # Remove NSCT numbers via regex\n",
    "    text = re.sub(r'nsct\\d*', ' ', text)\n",
    "    return text\n",
    "\n",
    "# Remove unnecessary text contents like punctuation, stopwords and text with no valuable information\n",
    "def transform_text(text):\n",
    "    text = text.lower()\n",
    "    text = removePunctuationAndSpecialCharacters(text)\n",
    "    text = removeSpecialWords(text)\n",
    "    words = nltk.word_tokenize(text, language=\"german\")\n",
    "    for word in words:\n",
    "        if word in stopwords.words('german'):\n",
    "            words.remove(word)\n",
    "        if word not in general_bag_of_words:\n",
    "            general_bag_of_words.append(word)\n",
    "    return \" \".join(words)\n",
    "\n",
    "# Vectorize a given text and count the occurence of words\n",
    "def vectorize(prepared_text):\n",
    "    vectorizer = CountVectorizer()\n",
    "    X = vectorizer.fit_transform(prepared_text).toarray()\n",
    "    topics = vectorizer.get_feature_names()\n",
    "    return pd.DataFrame(data=X, columns=topics)\n",
    "\n",
    "# Prepare basic data\n",
    "def prepare_data (dataset):\n",
    "    print(\"Preparing data...\")\n",
    "    prepared = dataset\n",
    "    prepared = prepared.fillna(0) # Fill NULL values with 0\n",
    "\n",
    "    #Transform columns to numerical values\n",
    "    prepared[columnname_account] = dataset[columnname_account].astype('category')\n",
    "    prepared[columnname_account] = prepared[columnname_account].cat.codes\n",
    "\n",
    "    prepared[columnname_blz] = dataset[columnname_blz].astype('category')\n",
    "    prepared[columnname_blz] = prepared[columnname_blz].cat.codes\n",
    "\n",
    "    prepared[columnname_text] = dataset[columnname_text].astype('category')\n",
    "    prepared[columnname_text] = prepared[columnname_text].cat.codes\n",
    "\n",
    "    prepared[columnname_depotnumber] = dataset[columnname_depotnumber].astype('category')\n",
    "    prepared[columnname_depotnumber] = prepared[columnname_depotnumber].cat.codes\n",
    "\n",
    "    prepared[columnname_currency] = dataset[columnname_currency].astype('category')\n",
    "    prepared[columnname_currency] = prepared[columnname_currency].cat.codes\n",
    "\n",
    "    prepared[columnname_orderdate] = pd.to_datetime(prepared[columnname_orderdate])\n",
    "    prepared[columnname_valutadate] = pd.to_datetime(prepared[columnname_valutadate])\n",
    "\n",
    "    # Normalize the \"Betrag\" feature via MinMax\n",
    "    prepared[columnname_amount] = prepared[columnname_amount].str.replace(',','.')\n",
    "    prepared[columnname_amount] = prepared[columnname_amount].astype('float')\n",
    "    mm_scaler = MinMaxScaler(feature_range=(0, 1))\n",
    "    prepared[columnname_amount] = mm_scaler.fit_transform(prepared[[columnname_amount]].values)\n",
    "\n",
    "    return prepared\n",
    "\n",
    "# Prepare the text features\n",
    "def prepare_text_data(dataset):\n",
    "    print(\"Preparing text featues...\")\n",
    "    usage = dataset[columnname_usage].apply(transform_text)\n",
    "    target_sender = dataset[columnname_target_sender].apply(transform_text)\n",
    "\n",
    "    dataframe_usage = vectorize(usage)\n",
    "    dataframe_targetsender = vectorize(target_sender)\n",
    "\n",
    "    return pd.concat([dataframe_usage, dataframe_targetsender], axis=1)\n",
    "\n",
    "# This will produce a complete prepared dataframe which can be used for training or also for prediction\n",
    "def run_complete_prepare(dataset):\n",
    "    print(\"Starting preparation of data...\")\n",
    "    prepared_data = prepare_data(dataset)\n",
    "    text_data = prepare_text_data(prepared_data)\n",
    "\n",
    "    # Remove text features as we use the results from the count vectorizer, also the Id is not required\n",
    "    prepared_data = prepared_data.drop([\"Unnamed: 0\", columnname_usage, columnname_target_sender], axis=1)\n",
    "\n",
    "    final_Data = pd.concat([prepared_data, text_data], axis=1)\n",
    "    final_Data = final_Data.loc[:,~final_Data.columns.duplicated()] # Remove duplicates from dataset\n",
    "    print(\"Prepared data!\")\n",
    "    return final_Data\n",
    "\n",
    "data_final = run_complete_prepare(data)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Prepared Data Statistics"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Preparing data...\n",
      "Preparing text featues...\n",
      "\n",
      "#################### Prepared Data #####################\n",
      "                            0  Null Values  count       mean        std  min  \\\n",
      "Auftragskonto            int8            0  209.0   0.377990   0.794024 -1.0   \n",
      "Buchungstag    datetime64[ns]            0    NaN        NaN        NaN  NaN   \n",
      "Valutadatum    datetime64[ns]            0    NaN        NaN        NaN  NaN   \n",
      "Buchungstext             int8            0  209.0   7.406699   2.784254  0.0   \n",
      "Kontonummer              int8            0  209.0  17.464115  16.175715 -1.0   \n",
      "BLZ                      int8            0  209.0  15.622010  14.501183 -1.0   \n",
      "Betrag                float64            0  209.0   0.626572   0.102435  0.0   \n",
      "Waehrung                 int8            0  209.0   0.000000   0.000000  0.0   \n",
      "label                  object            0    NaN        NaN        NaN  NaN   \n",
      "\n",
      "                    25%        50%        75%   max  \n",
      "Auftragskonto  0.000000   1.000000   1.000000   1.0  \n",
      "Buchungstag         NaN        NaN        NaN   NaN  \n",
      "Valutadatum         NaN        NaN        NaN   NaN  \n",
      "Buchungstext   7.000000   8.000000   9.000000  13.0  \n",
      "Kontonummer    2.000000  13.000000  31.000000  52.0  \n",
      "BLZ            2.000000  15.000000  27.000000  39.0  \n",
      "Betrag         0.613555   0.622464   0.624474   1.0  \n",
      "Waehrung       0.000000   0.000000   0.000000   0.0  \n",
      "label               NaN        NaN        NaN   NaN  \n",
      "\n",
      "################### Text Statistics ####################\n",
      "                 0  Null Values  count      mean       std  min  25%  50%  \\\n",
      "00           int64            0  209.0  0.253589  0.526051  0.0  0.0  0.0   \n",
      "0000000000   int64            0  209.0  0.162679  0.369959  0.0  0.0  0.0   \n",
      "001          int64            0  209.0  0.009569  0.097588  0.0  0.0  0.0   \n",
      "0084632      int64            0  209.0  0.004785  0.069171  0.0  0.0  0.0   \n",
      "00eur0       int64            0  209.0  0.100478  0.301359  0.0  0.0  0.0   \n",
      "...            ...          ...    ...       ...       ...  ...  ...  ...   \n",
      "vielen       int64            0  209.0  0.004785  0.069171  0.0  0.0  0.0   \n",
      "vr           int64            0  209.0  0.004785  0.069171  0.0  0.0  0.0   \n",
      "wuerzburger  int64            0  209.0  0.004785  0.069171  0.0  0.0  0.0   \n",
      "zalando      int64            0  209.0  0.014354  0.119231  0.0  0.0  0.0   \n",
      "zdf          int64            0  209.0  0.004785  0.069171  0.0  0.0  0.0   \n",
      "\n",
      "             75%  max  \n",
      "00           0.0  2.0  \n",
      "0000000000   0.0  1.0  \n",
      "001          0.0  1.0  \n",
      "0084632      0.0  1.0  \n",
      "00eur0       0.0  1.0  \n",
      "...          ...  ...  \n",
      "vielen       0.0  1.0  \n",
      "vr           0.0  1.0  \n",
      "wuerzburger  0.0  1.0  \n",
      "zalando      0.0  1.0  \n",
      "zdf          0.0  1.0  \n",
      "\n",
      "[472 rows x 10 columns]\n",
      "\n",
      "##################### Overall BoW ######################\n",
      "Length of Bow:  449\n",
      "['gehalt', 'adorsys', 'gmbh', 'co', 'kg', 'end', 'to', 'ref', 'kundenreferenz', 'gutschrift', 'byladem1sbt', 'de12773501123456789889', 'miete', 'beuthener', 'str', '25', 'dauerauftrag', '2107', '1634uhr', 'nuernberg', 'all', 'eur', '70,00', 'gebeur', '0,00', 'einzahlung', 'auszahlung', 'ga', 'nr55508568', 'blz76010085', '9', '2831', 'edeka', 'neubauer', 'nuernb', '2016', '07', '20t175445', 'kfn', 'vj', '1612', 'lastschrift', 'amazon', 'mietzuschuss', '05', '09t175445', 'dm', 'fil1137', 'h65128', 'de', '2', '016', '04', '19t170343folgenr001', 'verf', 'alld2020', '12', 'ccbade31xxx', 'de71310108339900123456', 'sparen', '1602', '10t175445', '02', '16t170343folgenr001', '04t175445', '06', '02t175445', '01t170343folgenr001', '0106', 'ssknde77xxx', 'de41760501010012345675', 'einmalsparen', 'ccb91ue162146', 'ueberweisung', '1904', 'spdude68xxx', '03', '15t175445', 'bears', '+', 'friends', 'gmb', 'verfa', 'lld2016', 'pulsnitz', 'elv56006915', '0804', '1535', 'me1rewe', 'sa', 'gt', 'danke', '46400517', 'tk', 'maxx', '201', '6', '02t135208', 'folgenr001', 'verfal', 'ld2020', 'visa62185021800', '279', '6620', '535,00eur0,0000000000', '3103', '10355999', '20', '16', '01', '02t141806', 'folgenr000', 'visa6218502106060876', '20,00eur0,0000000000', '0404', '20,00', '10316011', 'abonnement', 'kd', 'nr', '5104489', 'visa62185021senningerberg', '24,30eur0,0000000000', '0104', '10355499', '3,84eur0,0000000000', '12,99eur0,0000000000', '3003', '10355942', 'haushalt', 'urlaub', 'abschlag', 'strom', 'vk', '122721405', 'gruenstrom', 'classic', 'abschluss', 'bp1821187056', '3814', '50,00eur0,0000000000', '2503', '50,00', 'aramark', '24t123012', 'verfalld20', '230320211004393201252017040elv6520', '1704', '2303', '2021', 'm', 'e1', '2103', '1610', 'k358853507', 'rg', '150257651', '551', 'celona', '19t152252', 'verfall', 'd2020', '100,00eur0,0000000000', '1703', '100,00', '33,73eur0,0000000000', '1403', '14t120436', 'elv56006914', '1203', '1942', 'gas', '220200215414', 'elv66021447', '1730', 'me11', 'california', 't190455', 'verfalld2020', '40,00eur0,0000000000', '0903', '40,00', 'esso', 'deutschland', '09t070435', 've', 'rfalld2020', 'geschenk', 'doris', '04032016', '350193', 'horst', 'dietrich', '8324', 'reit', 'im', 'intersport', 'rent', '05t095727folgenr001', 'interne', 'umbuchung', '03t123304', 'rse', 'gastro', 'roethenbach', '02t220220', '0103', '1954', '2902', '52,42eur0,0000000000', '2702', 'visa62185021amazonde', '7,54eur0,0000000000', '2602', '10355969', 'omv', '7540', '26t151734', 'lld2020', 'mango', 'nuremberg', 'karolinen', 'nuernber', 'g', '0', '27t103917', 'folgenr', '001', 'verfalld2099', 'ticket', 'erg', '27t111418', 'fil2103', 'h65214', '27t114326folgenr001', '2502', '1951', 'visa62185021rathaus', '2302', '26,30eur0,0000000000', '1159', '7,99eur0,0000000000', '2202', '6,73eur0,0000000000', '10,00eur0,0000000000', '2102', 'gries', 'deco', 'company', '20t153251', 'ultra', 'comix', '20t173528', '150251073', '317', 'skyline', 'gaststaette', '18t213358folgenr001', 'visa62185021schnieglin', '140,00eur0,0000000000', '1302', '140,00', 'bil', '35,99eur0,0000000000', '1402', '10355735', '10,00', 'fuerthermare', 'fuerth', '13t', '143500', 'rundfunk', '012016', '032016', 'beitrags', '574704817', 'aenderungen', 'ganz', 'bequ', 'em', 'wwwrundfunkbeitragde', '1,01e+13', '0802', 're', '330000099057', '12272140', '5', 'cafe', 'bistro', 'seon', '05t230111', '0402', 'lombagine', 'apotheke', 'sebalder', 'h', '03t164807folgenr001', '161080109883418', '34439927', 'ts65354', 'travelsecure', 'auslandsrei', 'sekr', 'ankenvers', 'icherung', '310116', '3101', '17', '33,00', '303', '7483141', '2001136', 'mktplce', 'eu', '0164378887929309', '0580631', '8794708', '2411354884929964', '01t115514', '01t065722', '10103084736793zalando', '0084632', '5339566', '2410573888643087', 'elv56006917', '3001', '1537', '4488385', '6169917', '1637782719640441', 'vero', 'moda', '30t142554folgenr001', 'elv56006916', '2801', '1912', '2501', '1654', 'visa62185021nbg', 'koestr', '30,00eur0,0000000000', '2101', '30,00', 'n', 'style', 'nurnberg', '23t11', '3213', 'humanic', '23t13', '2317', 'euro', 'überweisung', '1901', 'visa62185013800', '19,95eur0,0000000000', '1801', '12,80eur0,0000000000', '150244534', '116', 'visa62185013nuernberg', '70,00eur0,0000000000', '1601', '10355311', 'visa62185013amazonde', '9,99eur0,0000000000', '1501', 'visa6218501306060876', '1401', '1301', '1101', '320000042615', '2202002154', '14', '0701', 'georg', 'tasche', 'bargeld', 'kartenzahlung', 'neue', 'playstation', 'isabel', 'anna', 'drogeriemarkt', 'sagt', 'michael', 'koffer', 'christina', 'jung', 'rewe', 'aeu', 'nürnberg', 'marketplace', 'bankhaus', 'august', 'lenz', 'atm', 'vag', 'foodora', 'fein', ',', 'lucas', 'stromio', 'billpay', 'ihs', 'teambank', 'sushi', 'glas', 'gastronomie', '1u1', 'telecom', 'bar', 'e', 'wie', 'h+m', '798', 'vielen', 'dank', 'hobex', 'ag', 'issuer', 'aus', 'sparkasse', 'gastst+veranst', 'digital', 'video', 'ard', 'zdf', 'dradio', 'beitragsservice', 'zalando', 'se', 'dankt', 'sarah', 'm¸ller', 'hoefe', 'wuerzburger', 'versicherungs', 'payments', 'europe', 'sca', '69', 'city', 'vr', 'bank', 'eg', 'breuninger', '*', 'deutsche']\n"
     ]
    }
   ],
   "source": [
    "######## Used for statistics ########\n",
    "data_prepared = prepare_data(data)\n",
    "data_text = prepare_text_data(data_prepared)\n",
    "# Remove text features as we use the results from the count vectorizer, also the Id is not required\n",
    "data_prepared = data_prepared.drop([\"Unnamed: 0\", columnname_usage, columnname_target_sender], axis=1)\n",
    "#####################################\n",
    "\n",
    "print(\"\\n#################### Prepared Data #####################\")\n",
    "data_prepared_summary = pd.DataFrame(data_prepared.dtypes)\n",
    "data_prepared_statistics = data_prepared.describe().transpose()\n",
    "data_prepared_summary[\"Null Values\"] = data_prepared.isnull().sum()\n",
    "data_prepared_summary = pd.concat([data_prepared_summary,data_prepared_statistics], axis=1)\n",
    "print(data_prepared_summary)\n",
    "\n",
    "print(\"\\n################### Text Statistics ####################\")\n",
    "data_text_summary = pd.DataFrame(data_text.dtypes)\n",
    "data_text_statistics = data_text.describe().transpose()\n",
    "data_text_summary[\"Null Values\"] = data_text.isnull().sum()\n",
    "data_text_summary = pd.concat([data_text_summary,data_text_statistics], axis=1)\n",
    "print(data_text_summary)\n",
    "\n",
    "print(\"\\n##################### Overall BoW ######################\")\n",
    "print(\"Length of Bow: \", len(general_bag_of_words))\n",
    "print(general_bag_of_words)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Feature Selection"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training|Test Size: 167|42\n"
     ]
    }
   ],
   "source": [
    "selected_features = data_final.drop([columnname_label, columnname_orderdate, columnname_valutadate], axis=1)\n",
    "selected_features_labels = data_final[columnname_label]\n",
    "train_data, test_data, train_labels, test_labels = train_test_split(selected_features, selected_features_labels, test_size=1-training_set_size, random_state=345)\n",
    "\n",
    "print(\"Training|Test Size: \" +str(len(train_data)) + \"|\"+str(len(test_data)))"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Naive Gaussian Classification\n",
    "## Training"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "outputs": [
    {
     "data": {
      "text/plain": "GaussianNB()"
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gnb_model = GaussianNB()\n",
    "gnb_model.fit(train_data, train_labels)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Evaluation\n",
    "### Accuracy and Class Precision"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy:             0.9285714285714286\n",
      "Null Accuracy:        0.2619047619047619\n",
      "\n",
      "\n",
      "Classification Report:   \n",
      "                   precision    recall  f1-score   support\n",
      "\n",
      "         finance       1.00      1.00      1.00         5\n",
      "          income       1.00      1.00      1.00         6\n",
      "         leisure       0.92      0.92      0.92        12\n",
      "          living       0.75      0.75      0.75         4\n",
      "         private       1.00      1.00      1.00         4\n",
      "standardOfLiving       0.91      0.91      0.91        11\n",
      "\n",
      "        accuracy                           0.93        42\n",
      "       macro avg       0.93      0.93      0.93        42\n",
      "    weighted avg       0.93      0.93      0.93        42\n",
      "\n",
      "Cross Validation Mean:  0.7863636363636364\n"
     ]
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "test_prediction = gnb_model.predict(test_data)\n",
    "\n",
    "accuracy = metrics.accuracy_score(test_labels,test_prediction)\n",
    "null_accuracy =11/len(test_labels)\n",
    "\n",
    "print(\"Accuracy:            \", accuracy)\n",
    "print(\"Null Accuracy:       \", null_accuracy)\n",
    "print(\"\\n\")\n",
    "print(\"Classification Report:   \\n\", metrics.classification_report(test_labels,test_prediction))\n",
    "\n",
    "cross_val_sc = cross_val_score(gnb_model, test_data, test_labels, cv=4)\n",
    "print(\"Cross Validation Mean: \", cross_val_sc.mean())"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Confusion Matrix"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "True Positives: 39\n",
      "False Positives: 2\n",
      "True Negatives: 1\n",
      "\n",
      "Classification Error:  0.07142857142857142\n"
     ]
    }
   ],
   "source": [
    "confusion_matrix_frame = pd.crosstab(test_labels, test_prediction, rownames=['Actual'], colnames=['Predicted'])\n",
    "\n",
    "def true_positives(matrix):\n",
    "    tp = 0\n",
    "    for i in range(6):\n",
    "        tp += matrix[i,i]\n",
    "    return tp\n",
    "\n",
    "def false_positives(matrix):\n",
    "    fp = 0\n",
    "    for i in range(6):\n",
    "        for y in range( i+1, 6, 1):\n",
    "            fp += matrix[i,y]\n",
    "    return fp\n",
    "\n",
    "def false_negatives(matrix):\n",
    "    fn = 0\n",
    "    for i in range(6):\n",
    "        for y in range( 0, i, 1):\n",
    "            fn += matrix[i,y]\n",
    "    return fn\n",
    "\n",
    "tmp_confusion_matrix = confusion_matrix(test_labels, test_prediction)\n",
    "\n",
    "tps = true_positives(tmp_confusion_matrix)\n",
    "fps = false_positives(tmp_confusion_matrix)\n",
    "fns = false_negatives(tmp_confusion_matrix)\n",
    "\n",
    "print(\"\\nTrue Positives: \" +str(tps))\n",
    "print(\"False Positives: \" +str(fps))\n",
    "print(\"True Negatives: \" +str(fns))\n",
    "\n",
    "classification_error = (fns +fps) / float(len(test_labels))\n",
    "print(\"\\nClassification Error: \", classification_error)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 288x288 with 2 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_confusion_matrix(df_confusion, title='Prediction Confusion Matrix', cmap=plt.cm.gray_r):\n",
    "    plt.matshow(df_confusion, cmap=cmap) # imshow\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(df_confusion.columns))\n",
    "    plt.xticks(tick_marks, df_confusion.columns, rotation=90)\n",
    "    plt.yticks(tick_marks, df_confusion.index)\n",
    "    plt.ylabel(df_confusion.index.name)\n",
    "    plt.xlabel(df_confusion.columns.name)\n",
    "\n",
    "plot_confusion_matrix(confusion_matrix_frame,cmap=plt.cm.hot_r)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix with values:   \n",
      " Predicted         finance  income  leisure  living  private  standardOfLiving\n",
      "Actual                                                                       \n",
      "finance                 5       0        0       0        0                 0\n",
      "income                  0       6        0       0        0                 0\n",
      "leisure                 0       0       11       1        0                 0\n",
      "living                  0       0        0       3        0                 1\n",
      "private                 0       0        0       0        4                 0\n",
      "standardOfLiving        0       0        1       0        0                10\n"
     ]
    }
   ],
   "source": [
    "print(\"Confusion Matrix with values:   \\n\", confusion_matrix_frame)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Prediction Probability Distribution"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_dis(class_proba):\n",
    "    plt.hist(class_proba, bins=8)\n",
    "    # x-axis limit from 0 to 1\n",
    "    plt.xlim(0,1)\n",
    "    plt.title('Histogram of predicted probabilities')\n",
    "    plt.xlabel('Predicted probability of classes')\n",
    "    plt.ylabel('Frequency')\n",
    "\n",
    "plot_dis(gnb_model.predict_proba(test_data))"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}