jupy_raf.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split,GridSearchCV,StratifiedShuffleSplit\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer,CountVectorizer\n",
    "from sklearn.linear_model import LogisticRegression,LogisticRegressionCV\n",
    "from sklearn.ensemble import GradientBoostingClassifier\n",
    "from sklearn.preprocessing import PolynomialFeatures\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.metrics import precision_recall_fscore_support,classification_report, confusion_matrix\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import numpy as np\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "class_names = ['suspicious','non suspicious']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "#These are the functions I will use for result visualisation\n",
    "def plot_confusion_matrix(y_test,y_predicted,labels):\n",
    "    cm = confusion_matrix(y_test, y_predicted,labels =labels)\n",
    "\n",
    "    figsize = (10,7)\n",
    "    df_cm = pd.DataFrame(\n",
    "        cm, index=class_names, columns=class_names, \n",
    "    )\n",
    "    fig = plt.figure(figsize=figsize)\n",
    "    try:\n",
    "        heatmap = sns.heatmap(df_cm, annot=True, fmt=\"d\")\n",
    "    except ValueError:\n",
    "        raise ValueError(\"Confusion matrix values must be integers.\")\n",
    "    heatmap.yaxis.set_ticklabels(heatmap.yaxis.get_ticklabels(), rotation=0)\n",
    "    heatmap.xaxis.set_ticklabels(heatmap.xaxis.get_ticklabels(), rotation=45)\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')\n",
    "    plt.show()\n",
    "    return\n",
    "\n",
    "def evaluate_classifier(pipeline, x_test, y_test):\n",
    "\n",
    "    y_predicted = pipeline.predict(x_test)\n",
    "\n",
    "    report  = classification_report(y_test,y_predicted)\n",
    "    print(report)\n",
    "    plot_confusion_matrix(y_test, y_predicted, class_names)\n",
    "    return"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 284,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv('./data/train.csv',index_col='customer',encoding='utf8')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 290,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    990913\n",
       "1      9087\n",
       "Name: suspicious, dtype: int64"
      ]
     },
     "execution_count": 290,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.suspicious.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 288,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 289,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'\"{\\\\\"90000000\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":734958.58,\\\\\"transaction_count\\\\\":7.0,\\\\\"io_ratio\\\\\":0.1428571429,\\\\\"age\\\\\":79.0,\\\\\"nationality\\\\\":134,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":24.63,\\\\\"inactive_days_max\\\\\":124.0,\\\\\"n_of_accounts\\\\\":3,\\\\\"distinct_counterparties\\\\\":6.0,\\\\\"channel_risk\\\\\":0.1287034449,\\\\\"atm_withdrawal\\\\\":215959.85,\\\\\"atm_deposit\\\\\":832.3648542642},\\\\\"90000001\\\\\":{\\\\\"category\\\\\":1,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":1009400.35,\\\\\"transaction_count\\\\\":2.0,\\\\\"io_ratio\\\\\":0.5,\\\\\"age\\\\\":0.0,\\\\\"nationality\\\\\":33,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":78.49,\\\\\"inactive_days_max\\\\\":79.0,\\\\\"n_of_accounts\\\\\":3,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.6355841857,\\\\\"atm_withdrawal\\\\\":53.97,\\\\\"atm_deposit\\\\\":34.7522980256},\\\\\"90000002\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":11051.68,\\\\\"transaction_count\\\\\":3.0,\\\\\"io_ratio\\\\\":0.6666666667,\\\\\"age\\\\\":77.0,\\\\\"nationality\\\\\":191,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":35.74,\\\\\"inactive_days_max\\\\\":72.0,\\\\\"n_of_accounts\\\\\":2,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.1381515263,\\\\\"atm_withdrawal\\\\\":1632.91,\\\\\"atm_deposit\\\\\":17.3230309692},\\\\\"90000003\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":31193.29,\\\\\"transaction_count\\\\\":81.0,\\\\\"io_ratio\\\\\":0.3086419753,\\\\\"age\\\\\":77.0,\\\\\"nationality\\\\\":50,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":2.05,\\\\\"inactive_days_max\\\\\":95.0,\\\\\"n_of_accounts\\\\\":2,\\\\\"distinct_counterparties\\\\\":2.0,\\\\\"channel_risk\\\\\":0.3339038401,\\\\\"atm_withdrawal\\\\\":3203.1,\\\\\"atm_deposit\\\\\":50.9452806876},\\\\\"90000004\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":205489.77,\\\\\"transaction_count\\\\\":1.0,\\\\\"io_ratio\\\\\":1.0,\\\\\"age\\\\\":39.0,\\\\\"nationality\\\\\":172,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":147.53,\\\\\"inactive_days_max\\\\\":148.0,\\\\\"n_of_accounts\\\\\":4,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.3415776601,\\\\\"atm_withdrawal\\\\\":5626.65,\\\\\"atm_deposit\\\\\":0.0},\\\\\"90000005\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":279.73,\\\\\"transaction_count\\\\\":3.0,\\\\\"io_ratio\\\\\":0.3333333333,\\\\\"age\\\\\":36.0,\\\\\"nationality\\\\\":172,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":44.6,\\\\\"inactive_days_max\\\\\":90.0,\\\\\"n_of_accounts\\\\\":2,\\\\\"distinct_counterparties\\\\\":3.0,\\\\\"channel_risk\\\\\":0.3281297198,\\\\\"atm_withdrawal\\\\\":41.38,\\\\\"atm_deposit\\\\\":0.5414054227},\\\\\"90000006\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":27879.68,\\\\\"transaction_count\\\\\":1.0,\\\\\"io_ratio\\\\\":0.0,\\\\\"age\\\\\":37.0,\\\\\"nationality\\\\\":45,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":89.55,\\\\\"inactive_days_max\\\\\":180.0,\\\\\"n_of_accounts\\\\\":2,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.1419474808,\\\\\"atm_withdrawal\\\\\":0.0,\\\\\"atm_deposit\\\\\":55.061748457},\\\\\"90000007\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":192441.04,\\\\\"transaction_count\\\\\":105.0,\\\\\"io_ratio\\\\\":0.3523809524,\\\\\"age\\\\\":61.0,\\\\\"nationality\\\\\":172,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":1.83,\\\\\"inactive_days_max\\\\\":116.0,\\\\\"n_of_accounts\\\\\":2,\\\\\"distinct_counterparties\\\\\":5.0,\\\\\"channel_risk\\\\\":0.4407882001,\\\\\"atm_withdrawal\\\\\":6486.87,\\\\\"atm_deposit\\\\\":11.6133255649},\\\\\"90000008\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":55066.73,\\\\\"transaction_count\\\\\":2.0,\\\\\"io_ratio\\\\\":0.5,\\\\\"age\\\\\":31.0,\\\\\"nationality\\\\\":172,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":69.76,\\\\\"inactive_days_max\\\\\":140.0,\\\\\"n_of_accounts\\\\\":2,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.1815844131,\\\\\"atm_withdrawal\\\\\":5853.38,\\\\\"atm_deposit\\\\\":344.853306209},\\\\\"90000009\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":39803.45,\\\\\"transaction_count\\\\\":1.0,\\\\\"io_ratio\\\\\":0.0,\\\\\"age\\\\\":59.0,\\\\\"nationality\\\\\":33,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":140.33,\\\\\"inactive_days_max\\\\\":141.0,\\\\\"n_of_accounts\\\\\":3,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.2279588376,\\\\\"atm_withdrawal\\\\\":39803.45,\\\\\"atm_deposit\\\\\":43.5873256829},\\\\\"90000010\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":94542.7,\\\\\"transaction_count\\\\\":3.0,\\\\\"io_ratio\\\\\":0.3333333333,\\\\\"age\\\\\":57.0,\\\\\"nationality\\\\\":50,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":81.55,\\\\\"inactive_days_max\\\\\":327.0,\\\\\"n_of_accounts\\\\\":1,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.1459706837,\\\\\"atm_withdrawal\\\\\":3549.54,\\\\\"atm_deposit\\\\\":343.2800871096},\\\\\"90000011\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":222229.46,\\\\\"transaction_count\\\\\":1.0,\\\\\"io_ratio\\\\\":0.0,\\\\\"age\\\\\":57.0,\\\\\"nationality\\\\\":33,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":18.65,\\\\\"inactive_days_max\\\\\":19.0,\\\\\"n_of_accounts\\\\\":3,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.0174927825,\\\\\"atm_withdrawal\\\\\":222229.46,\\\\\"atm_deposit\\\\\":3.3757004547},\\\\\"90000012\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":135202.54,\\\\\"transaction_count\\\\\":1.0,\\\\\"io_ratio\\\\\":0.0,\\\\\"age\\\\\":11.0,\\\\\"nationality\\\\\":94,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":28.19,\\\\\"inactive_days_max\\\\\":29.0,\\\\\"n_of_accounts\\\\\":1,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.3186248671,\\\\\"atm_withdrawal\\\\\":135202.54,\\\\\"atm_deposit\\\\\":32.9380388394},\\\\\"90000013\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":1283833.28,\\\\\"transaction_count\\\\\":2.0,\\\\\"io_ratio\\\\\":0.0,\\\\\"age\\\\\":23.0,\\\\\"nationality\\\\\":32,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":31.78,\\\\\"inactive_days_max\\\\\":64.0,\\\\\"n_of_accounts\\\\\":1,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.2315505238,\\\\\"atm_withdrawal\\\\\":1283833.28,\\\\\"atm_deposit\\\\\":809.256477516},\\\\\"90000014\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":23459.44,\\\\\"transaction_count\\\\\":25.0,\\\\\"io_ratio\\\\\":0.28,\\\\\"age\\\\\":35.0,\\\\\"nationality\\\\\":172,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":8.66,\\\\\"inactive_days_max\\\\\":113.0,\\\\\"n_of_accounts\\\\\":1,\\\\\"distinct_counterparties\\\\\":9.0,\\\\\"channel_risk\\\\\":0.3928379335,\\\\\"atm_withdrawal\\\\\":1280.56,\\\\\"atm_deposit\\\\\":4.2044768528},\\\\\"90000015\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":158174.63,\\\\\"transaction_count\\\\\":1.0,\\\\\"io_ratio\\\\\":1.0,\\\\\"age\\\\\":54.0,\\\\\"nationality\\\\\":172,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":102.52,\\\\\"inactive_days_max\\\\\":103.0,\\\\\"n_of_accounts\\\\\":2,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.1779819069,\\\\\"atm_withdrawal\\\\\":5196.25,\\\\\"atm_deposit\\\\\":0.0},\\\\\"90000016\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":22913.19,\\\\\"transaction_count\\\\\":12.0,\\\\\"io_ratio\\\\\":0.3333333333,\\\\\"age\\\\\":35.0,\\\\\"nationality\\\\\":94,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":16.9,\\\\\"inactive_days_max\\\\\":85.0,\\\\\"n_of_accounts\\\\\":3,\\\\\"distinct_counterparties\\\\\":2.0,\\\\\"channel_risk\\\\\":0.5327577074,\\\\\"atm_withdrawal\\\\\":2329.13,\\\\\"atm_deposit\\\\\":10.8279680038},\\\\\"90000017\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":1059637.26,\\\\\"transaction_count\\\\\":1.0,\\\\\"io_ratio\\\\\":1.0,\\\\\"age\\\\\":65.0,\\\\\"nationality\\\\\":33,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":87.48,\\\\\"inactive_days_max\\\\\":175.0,\\\\\"n_of_accounts\\\\\":3,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.0838716803,\\\\\"atm_withdrawal\\\\\":58955.87,\\\\\"atm_deposit\\\\\":0.0},\\\\\"90000018\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":88738.94,\\\\\"transaction_count\\\\\":1.0,\\\\\"io_ratio\\\\\":0.0,\\\\\"age\\\\\":14.0,\\\\\"nationality\\\\\":94,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":8.72,\\\\\"inactive_days_max\\\\\":9.0,\\\\\"n_of_accounts\\\\\":2,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.236291242,\\\\\"atm_withdrawal\\\\\":88738.94,\\\\\"atm_deposit\\\\\":1511.2792659419},\\\\\"90000019\\\\\":{\\\\\"category\\\\\":0,\\\\\"suspicious\\\\\":0,\\\\\"turnover\\\\\":361597.34,\\\\\"transaction_count\\\\\":2.0,\\\\\"io_ratio\\\\\":0.5,\\\\\"age\\\\\":72.0,\\\\\"nationality\\\\\":172,\\\\\"is_pep\\\\\":0,\\\\\"inactive_days_average\\\\\":45.44,\\\\\"inactive_days_max\\\\\":46.0,\\\\\"n_of_accounts\\\\\":3,\\\\\"distinct_counterparties\\\\\":1.0,\\\\\"channel_risk\\\\\":0.6802505797,\\\\\"atm_withdrawal\\\\\":6932.87,\\\\\"atm_deposit\\\\\":1305.1087057721}}\"'"
      ]
     },
     "execution_count": 289,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "json.dumps(data[0:][:20].to_json(orient='index'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_test = pd.read_csv('./data/test.csv',index_col='customer',encoding='utf8')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "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>turnover</th>\n",
       "      <th>transaction_count</th>\n",
       "      <th>io_ratio</th>\n",
       "      <th>age</th>\n",
       "      <th>nationality</th>\n",
       "      <th>is_pep</th>\n",
       "      <th>inactive_days_average</th>\n",
       "      <th>inactive_days_max</th>\n",
       "      <th>n_of_accounts</th>\n",
       "      <th>distinct_counterparties</th>\n",
       "      <th>channel_risk</th>\n",
       "      <th>atm_withdrawal</th>\n",
       "      <th>atm_deposit</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer</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",
       "      <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>91000000</th>\n",
       "      <td>0</td>\n",
       "      <td>1235859.85</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>64.0</td>\n",
       "      <td>33</td>\n",
       "      <td>0</td>\n",
       "      <td>106.42</td>\n",
       "      <td>107.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.046108</td>\n",
       "      <td>1235859.85</td>\n",
       "      <td>12549.438865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91000001</th>\n",
       "      <td>1</td>\n",
       "      <td>53087.97</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>172</td>\n",
       "      <td>0</td>\n",
       "      <td>69.28</td>\n",
       "      <td>139.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.414826</td>\n",
       "      <td>53087.97</td>\n",
       "      <td>116.106050</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91000002</th>\n",
       "      <td>0</td>\n",
       "      <td>22003.25</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.25</td>\n",
       "      <td>41.0</td>\n",
       "      <td>33</td>\n",
       "      <td>0</td>\n",
       "      <td>30.94</td>\n",
       "      <td>62.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.030105</td>\n",
       "      <td>1066.98</td>\n",
       "      <td>106.434228</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91000003</th>\n",
       "      <td>0</td>\n",
       "      <td>72947.07</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.50</td>\n",
       "      <td>16.0</td>\n",
       "      <td>94</td>\n",
       "      <td>0</td>\n",
       "      <td>24.19</td>\n",
       "      <td>97.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.025249</td>\n",
       "      <td>16701.34</td>\n",
       "      <td>106.455505</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91000004</th>\n",
       "      <td>0</td>\n",
       "      <td>9964.06</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>46.0</td>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>47.84</td>\n",
       "      <td>48.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.083395</td>\n",
       "      <td>9964.06</td>\n",
       "      <td>118.541699</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          category    turnover  transaction_count  io_ratio   age  \\\n",
       "customer                                                            \n",
       "91000000         0  1235859.85                1.0      0.00  64.0   \n",
       "91000001         1    53087.97                3.0      0.00   0.0   \n",
       "91000002         0    22003.25                4.0      0.25  41.0   \n",
       "91000003         0    72947.07                4.0      0.50  16.0   \n",
       "91000004         0     9964.06                1.0      0.00  46.0   \n",
       "\n",
       "          nationality  is_pep  inactive_days_average  inactive_days_max  \\\n",
       "customer                                                                  \n",
       "91000000           33       0                 106.42              107.0   \n",
       "91000001          172       0                  69.28              139.0   \n",
       "91000002           33       0                  30.94               62.0   \n",
       "91000003           94       0                  24.19               97.0   \n",
       "91000004           50       0                  47.84               48.0   \n",
       "\n",
       "          n_of_accounts  distinct_counterparties  channel_risk  \\\n",
       "customer                                                         \n",
       "91000000              2                      1.0      0.046108   \n",
       "91000001              1                      1.0      0.414826   \n",
       "91000002              2                      1.0      0.030105   \n",
       "91000003              2                      1.0      0.025249   \n",
       "91000004              2                      1.0      0.083395   \n",
       "\n",
       "          atm_withdrawal   atm_deposit  \n",
       "customer                                \n",
       "91000000      1235859.85  12549.438865  \n",
       "91000001        53087.97    116.106050  \n",
       "91000002         1066.98    106.434228  \n",
       "91000003        16701.34    106.455505  \n",
       "91000004         9964.06    118.541699  "
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "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>suspicious</th>\n",
       "      <th>turnover</th>\n",
       "      <th>transaction_count</th>\n",
       "      <th>io_ratio</th>\n",
       "      <th>age</th>\n",
       "      <th>nationality</th>\n",
       "      <th>is_pep</th>\n",
       "      <th>inactive_days_average</th>\n",
       "      <th>inactive_days_max</th>\n",
       "      <th>n_of_accounts</th>\n",
       "      <th>distinct_counterparties</th>\n",
       "      <th>channel_risk</th>\n",
       "      <th>atm_withdrawal</th>\n",
       "      <th>atm_deposit</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer</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",
       "      <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>90000000</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>734958.58</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>79.0</td>\n",
       "      <td>134</td>\n",
       "      <td>0</td>\n",
       "      <td>24.63</td>\n",
       "      <td>124.0</td>\n",
       "      <td>3</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.128703</td>\n",
       "      <td>215959.85</td>\n",
       "      <td>832.364854</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000001</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1009400.35</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>33</td>\n",
       "      <td>0</td>\n",
       "      <td>78.49</td>\n",
       "      <td>79.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.635584</td>\n",
       "      <td>53.97</td>\n",
       "      <td>34.752298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000002</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>11051.68</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>77.0</td>\n",
       "      <td>191</td>\n",
       "      <td>0</td>\n",
       "      <td>35.74</td>\n",
       "      <td>72.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.138152</td>\n",
       "      <td>1632.91</td>\n",
       "      <td>17.323031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000003</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>31193.29</td>\n",
       "      <td>81.0</td>\n",
       "      <td>0.308642</td>\n",
       "      <td>77.0</td>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>2.05</td>\n",
       "      <td>95.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.333904</td>\n",
       "      <td>3203.10</td>\n",
       "      <td>50.945281</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000004</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>205489.77</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>39.0</td>\n",
       "      <td>172</td>\n",
       "      <td>0</td>\n",
       "      <td>147.53</td>\n",
       "      <td>148.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.341578</td>\n",
       "      <td>5626.65</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          category  suspicious    turnover  transaction_count  io_ratio   age  \\\n",
       "customer                                                                        \n",
       "90000000         0           0   734958.58                7.0  0.142857  79.0   \n",
       "90000001         1           0  1009400.35                2.0  0.500000   0.0   \n",
       "90000002         0           0    11051.68                3.0  0.666667  77.0   \n",
       "90000003         0           0    31193.29               81.0  0.308642  77.0   \n",
       "90000004         0           0   205489.77                1.0  1.000000  39.0   \n",
       "\n",
       "          nationality  is_pep  inactive_days_average  inactive_days_max  \\\n",
       "customer                                                                  \n",
       "90000000          134       0                  24.63              124.0   \n",
       "90000001           33       0                  78.49               79.0   \n",
       "90000002          191       0                  35.74               72.0   \n",
       "90000003           50       0                   2.05               95.0   \n",
       "90000004          172       0                 147.53              148.0   \n",
       "\n",
       "          n_of_accounts  distinct_counterparties  channel_risk  \\\n",
       "customer                                                         \n",
       "90000000              3                      6.0      0.128703   \n",
       "90000001              3                      1.0      0.635584   \n",
       "90000002              2                      1.0      0.138152   \n",
       "90000003              2                      2.0      0.333904   \n",
       "90000004              4                      1.0      0.341578   \n",
       "\n",
       "          atm_withdrawal  atm_deposit  \n",
       "customer                               \n",
       "90000000       215959.85   832.364854  \n",
       "90000001           53.97    34.752298  \n",
       "90000002         1632.91    17.323031  \n",
       "90000003         3203.10    50.945281  \n",
       "90000004         5626.65     0.000000  "
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols_to_transform = ['nationality', 'is_pep','category']\n",
    "data = pd.get_dummies(data=data, columns = cols_to_transform)\n",
    "data['suspicious'] = data['suspicious'].map({ 1 : 'suspicious', 0: 'non suspicious'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols_to_transform = ['nationality', 'is_pep','category']\n",
    "data_test = pd.get_dummies(data=data_test, columns = cols_to_transform)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "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>turnover</th>\n",
       "      <th>transaction_count</th>\n",
       "      <th>io_ratio</th>\n",
       "      <th>age</th>\n",
       "      <th>inactive_days_average</th>\n",
       "      <th>inactive_days_max</th>\n",
       "      <th>n_of_accounts</th>\n",
       "      <th>distinct_counterparties</th>\n",
       "      <th>channel_risk</th>\n",
       "      <th>atm_withdrawal</th>\n",
       "      <th>...</th>\n",
       "      <th>nationality_134</th>\n",
       "      <th>nationality_141</th>\n",
       "      <th>nationality_172</th>\n",
       "      <th>nationality_176</th>\n",
       "      <th>nationality_191</th>\n",
       "      <th>is_pep_0</th>\n",
       "      <th>is_pep_1</th>\n",
       "      <th>category_0</th>\n",
       "      <th>category_1</th>\n",
       "      <th>category_2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer</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",
       "      <th></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",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>90000000</th>\n",
       "      <td>734958.58</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>79.0</td>\n",
       "      <td>24.63</td>\n",
       "      <td>124.0</td>\n",
       "      <td>3</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.128703</td>\n",
       "      <td>215959.85</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000001</th>\n",
       "      <td>1009400.35</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>78.49</td>\n",
       "      <td>79.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.635584</td>\n",
       "      <td>53.97</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000002</th>\n",
       "      <td>11051.68</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>77.0</td>\n",
       "      <td>35.74</td>\n",
       "      <td>72.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.138152</td>\n",
       "      <td>1632.91</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000003</th>\n",
       "      <td>31193.29</td>\n",
       "      <td>81.0</td>\n",
       "      <td>0.308642</td>\n",
       "      <td>77.0</td>\n",
       "      <td>2.05</td>\n",
       "      <td>95.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.333904</td>\n",
       "      <td>3203.10</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000004</th>\n",
       "      <td>205489.77</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>39.0</td>\n",
       "      <td>147.53</td>\n",
       "      <td>148.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.341578</td>\n",
       "      <td>5626.65</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 32 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            turnover  transaction_count  io_ratio   age  \\\n",
       "customer                                                  \n",
       "90000000   734958.58                7.0  0.142857  79.0   \n",
       "90000001  1009400.35                2.0  0.500000   0.0   \n",
       "90000002    11051.68                3.0  0.666667  77.0   \n",
       "90000003    31193.29               81.0  0.308642  77.0   \n",
       "90000004   205489.77                1.0  1.000000  39.0   \n",
       "\n",
       "          inactive_days_average  inactive_days_max  n_of_accounts  \\\n",
       "customer                                                            \n",
       "90000000                  24.63              124.0              3   \n",
       "90000001                  78.49               79.0              3   \n",
       "90000002                  35.74               72.0              2   \n",
       "90000003                   2.05               95.0              2   \n",
       "90000004                 147.53              148.0              4   \n",
       "\n",
       "          distinct_counterparties  channel_risk  atm_withdrawal     ...      \\\n",
       "customer                                                            ...       \n",
       "90000000                      6.0      0.128703       215959.85     ...       \n",
       "90000001                      1.0      0.635584           53.97     ...       \n",
       "90000002                      1.0      0.138152         1632.91     ...       \n",
       "90000003                      2.0      0.333904         3203.10     ...       \n",
       "90000004                      1.0      0.341578         5626.65     ...       \n",
       "\n",
       "          nationality_134  nationality_141  nationality_172  nationality_176  \\\n",
       "customer                                                                       \n",
       "90000000                1                0                0                0   \n",
       "90000001                0                0                0                0   \n",
       "90000002                0                0                0                0   \n",
       "90000003                0                0                0                0   \n",
       "90000004                0                0                1                0   \n",
       "\n",
       "          nationality_191  is_pep_0  is_pep_1  category_0  category_1  \\\n",
       "customer                                                                \n",
       "90000000                0         1         0           1           0   \n",
       "90000001                0         1         0           0           1   \n",
       "90000002                1         1         0           1           0   \n",
       "90000003                0         1         0           1           0   \n",
       "90000004                0         1         0           1           0   \n",
       "\n",
       "          category_2  \n",
       "customer              \n",
       "90000000           0  \n",
       "90000001           0  \n",
       "90000002           0  \n",
       "90000003           0  \n",
       "90000004           0  \n",
       "\n",
       "[5 rows x 32 columns]"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_test['nationality_117'] = 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "See label frequencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_frequencies = []\n",
    "label_frequencies.append(data.suspicious.tolist().count('suspicious')/len(data))\n",
    "label_frequencies.append(data.suspicious.tolist().count('non suspicious')/len(data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.009087, 0.990913]"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_frequencies"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "See nans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                         Nb of null obs  Percentage\n",
      "category                              0         0.0\n",
      "suspicious                            0         0.0\n",
      "turnover                              0         0.0\n",
      "transaction_count                     0         0.0\n",
      "io_ratio                              0         0.0\n",
      "age                                   0         0.0\n",
      "nationality                           0         0.0\n",
      "is_pep                                0         0.0\n",
      "inactive_days_average                 0         0.0\n",
      "inactive_days_max                     0         0.0\n",
      "n_of_accounts                         0         0.0\n",
      "distinct_counterparties               0         0.0\n",
      "channel_risk                          0         0.0\n",
      "atm_withdrawal                        0         0.0\n",
      "atm_deposit                           0         0.0\n"
     ]
    }
   ],
   "source": [
    "# Compute the number of missing observations for each feature\n",
    "nb_null = data.apply(lambda x: x.isnull().sum(), axis=0).nlargest(20)\n",
    "\n",
    "# Compute the percentage of missing values for each feature\n",
    "pct_null = nb_null/len(data)\n",
    "\n",
    "# Create a new DataFrame to display the results\n",
    "d = {'Nb of null obs':nb_null, 'Percentage': pct_null}\n",
    "df = pd.DataFrame(d)\n",
    "\n",
    "print(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>category</th>\n",
       "      <th>suspicious</th>\n",
       "      <th>turnover</th>\n",
       "      <th>transaction_count</th>\n",
       "      <th>io_ratio</th>\n",
       "      <th>age</th>\n",
       "      <th>nationality</th>\n",
       "      <th>is_pep</th>\n",
       "      <th>inactive_days_average</th>\n",
       "      <th>inactive_days_max</th>\n",
       "      <th>n_of_accounts</th>\n",
       "      <th>distinct_counterparties</th>\n",
       "      <th>channel_risk</th>\n",
       "      <th>atm_withdrawal</th>\n",
       "      <th>atm_deposit</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer</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",
       "      <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>90000000</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>734958.58</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>79.0</td>\n",
       "      <td>134</td>\n",
       "      <td>0</td>\n",
       "      <td>24.63</td>\n",
       "      <td>124.0</td>\n",
       "      <td>3</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.128703</td>\n",
       "      <td>215959.85</td>\n",
       "      <td>832.364854</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000001</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1009400.35</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>33</td>\n",
       "      <td>0</td>\n",
       "      <td>78.49</td>\n",
       "      <td>79.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.635584</td>\n",
       "      <td>53.97</td>\n",
       "      <td>34.752298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000002</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>11051.68</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>77.0</td>\n",
       "      <td>191</td>\n",
       "      <td>0</td>\n",
       "      <td>35.74</td>\n",
       "      <td>72.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.138152</td>\n",
       "      <td>1632.91</td>\n",
       "      <td>17.323031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000003</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>31193.29</td>\n",
       "      <td>81.0</td>\n",
       "      <td>0.308642</td>\n",
       "      <td>77.0</td>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "      <td>2.05</td>\n",
       "      <td>95.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.333904</td>\n",
       "      <td>3203.10</td>\n",
       "      <td>50.945281</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000004</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>205489.77</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>39.0</td>\n",
       "      <td>172</td>\n",
       "      <td>0</td>\n",
       "      <td>147.53</td>\n",
       "      <td>148.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.341578</td>\n",
       "      <td>5626.65</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          category  suspicious    turnover  transaction_count  io_ratio   age  \\\n",
       "customer                                                                        \n",
       "90000000         0           0   734958.58                7.0  0.142857  79.0   \n",
       "90000001         1           0  1009400.35                2.0  0.500000   0.0   \n",
       "90000002         0           0    11051.68                3.0  0.666667  77.0   \n",
       "90000003         0           0    31193.29               81.0  0.308642  77.0   \n",
       "90000004         0           0   205489.77                1.0  1.000000  39.0   \n",
       "\n",
       "          nationality  is_pep  inactive_days_average  inactive_days_max  \\\n",
       "customer                                                                  \n",
       "90000000          134       0                  24.63              124.0   \n",
       "90000001           33       0                  78.49               79.0   \n",
       "90000002          191       0                  35.74               72.0   \n",
       "90000003           50       0                   2.05               95.0   \n",
       "90000004          172       0                 147.53              148.0   \n",
       "\n",
       "          n_of_accounts  distinct_counterparties  channel_risk  \\\n",
       "customer                                                         \n",
       "90000000              3                      6.0      0.128703   \n",
       "90000001              3                      1.0      0.635584   \n",
       "90000002              2                      1.0      0.138152   \n",
       "90000003              2                      2.0      0.333904   \n",
       "90000004              4                      1.0      0.341578   \n",
       "\n",
       "          atm_withdrawal  atm_deposit  \n",
       "customer                               \n",
       "90000000       215959.85   832.364854  \n",
       "90000001           53.97    34.752298  \n",
       "90000002         1632.91    17.323031  \n",
       "90000003         3203.10    50.945281  \n",
       "90000004         5626.65     0.000000  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = data.suspicious\n",
    "x = data.drop('suspicious',axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_x, test_x, train_y, test_y = train_test_split(x,y,test_size=0.2, random_state= 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "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>turnover</th>\n",
       "      <th>transaction_count</th>\n",
       "      <th>io_ratio</th>\n",
       "      <th>age</th>\n",
       "      <th>inactive_days_average</th>\n",
       "      <th>inactive_days_max</th>\n",
       "      <th>n_of_accounts</th>\n",
       "      <th>distinct_counterparties</th>\n",
       "      <th>channel_risk</th>\n",
       "      <th>atm_withdrawal</th>\n",
       "      <th>...</th>\n",
       "      <th>nationality_134</th>\n",
       "      <th>nationality_141</th>\n",
       "      <th>nationality_172</th>\n",
       "      <th>nationality_176</th>\n",
       "      <th>nationality_191</th>\n",
       "      <th>is_pep_0</th>\n",
       "      <th>is_pep_1</th>\n",
       "      <th>category_0</th>\n",
       "      <th>category_1</th>\n",
       "      <th>category_2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer</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",
       "      <th></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",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>90771718</th>\n",
       "      <td>409868.95</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>60.0</td>\n",
       "      <td>30.46</td>\n",
       "      <td>92.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.410764</td>\n",
       "      <td>49313.47</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90521462</th>\n",
       "      <td>143538.46</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>26.0</td>\n",
       "      <td>94.21</td>\n",
       "      <td>189.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.184895</td>\n",
       "      <td>0.00</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90137361</th>\n",
       "      <td>770447.14</td>\n",
       "      <td>26.0</td>\n",
       "      <td>0.269231</td>\n",
       "      <td>48.0</td>\n",
       "      <td>6.11</td>\n",
       "      <td>98.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.130168</td>\n",
       "      <td>4727.42</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90404985</th>\n",
       "      <td>2407.86</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>18.0</td>\n",
       "      <td>49.89</td>\n",
       "      <td>150.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.216123</td>\n",
       "      <td>683.03</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90910092</th>\n",
       "      <td>1093644.44</td>\n",
       "      <td>18.0</td>\n",
       "      <td>0.555556</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7.87</td>\n",
       "      <td>79.0</td>\n",
       "      <td>5</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.607511</td>\n",
       "      <td>117031.88</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 32 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            turnover  transaction_count  io_ratio   age  \\\n",
       "customer                                                  \n",
       "90771718   409868.95                3.0  0.666667  60.0   \n",
       "90521462   143538.46                3.0  0.000000  26.0   \n",
       "90137361   770447.14               26.0  0.269231  48.0   \n",
       "90404985     2407.86                4.0  0.250000  18.0   \n",
       "90910092  1093644.44               18.0  0.555556   0.0   \n",
       "\n",
       "          inactive_days_average  inactive_days_max  n_of_accounts  \\\n",
       "customer                                                            \n",
       "90771718                  30.46               92.0              1   \n",
       "90521462                  94.21              189.0              1   \n",
       "90137361                   6.11               98.0              1   \n",
       "90404985                  49.89              150.0              1   \n",
       "90910092                   7.87               79.0              5   \n",
       "\n",
       "          distinct_counterparties  channel_risk  atm_withdrawal     ...      \\\n",
       "customer                                                            ...       \n",
       "90771718                      1.0      0.410764        49313.47     ...       \n",
       "90521462                      1.0      0.184895            0.00     ...       \n",
       "90137361                      1.0      0.130168         4727.42     ...       \n",
       "90404985                      1.0      0.216123          683.03     ...       \n",
       "90910092                      1.0      0.607511       117031.88     ...       \n",
       "\n",
       "          nationality_134  nationality_141  nationality_172  nationality_176  \\\n",
       "customer                                                                       \n",
       "90771718                0                0                1                0   \n",
       "90521462                0                0                1                0   \n",
       "90137361                0                0                0                0   \n",
       "90404985                0                0                0                0   \n",
       "90910092                0                0                1                0   \n",
       "\n",
       "          nationality_191  is_pep_0  is_pep_1  category_0  category_1  \\\n",
       "customer                                                                \n",
       "90771718                0         1         0           1           0   \n",
       "90521462                0         1         0           1           0   \n",
       "90137361                0         1         0           1           0   \n",
       "90404985                0         1         0           1           0   \n",
       "90910092                0         1         0           0           1   \n",
       "\n",
       "          category_2  \n",
       "customer              \n",
       "90771718           0  \n",
       "90521462           0  \n",
       "90137361           0  \n",
       "90404985           0  \n",
       "90910092           0  \n",
       "\n",
       "[5 rows x 32 columns]"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_x.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline = Pipeline([\n",
    "    ('classifier',          LogisticRegression(class_weight='balanced',n_jobs=-1))\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Raphael\\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\\Raphael\\Anaconda3\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:1296: UserWarning: 'n_jobs' > 1 does not have any effect when 'solver' is set to 'liblinear'. Got 'n_jobs' = 8.\n",
      "  \" = {}.\".format(effective_n_jobs(self.n_jobs)))\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "     steps=[('classifier', LogisticRegression(C=1.0, class_weight='balanced', dual=False,\n",
       "          fit_intercept=True, intercept_scaling=1, max_iter=100,\n",
       "          multi_class='warn', n_jobs=-1, penalty='l2', random_state=None,\n",
       "          solver='warn', tol=0.0001, verbose=0, warm_start=False))])"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipeline.fit(train_x,train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.99      0.45      0.62    198157\n",
      "           1       0.01      0.62      0.02      1843\n",
      "\n",
      "   micro avg       0.46      0.46      0.46    200000\n",
      "   macro avg       0.50      0.54      0.32    200000\n",
      "weighted avg       0.98      0.46      0.62    200000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "evaluate_classifier(pipeline, test_x,test_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "sss = StratifiedShuffleSplit(n_splits=2, test_size=0.2, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "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>turnover</th>\n",
       "      <th>transaction_count</th>\n",
       "      <th>io_ratio</th>\n",
       "      <th>age</th>\n",
       "      <th>inactive_days_average</th>\n",
       "      <th>inactive_days_max</th>\n",
       "      <th>n_of_accounts</th>\n",
       "      <th>distinct_counterparties</th>\n",
       "      <th>channel_risk</th>\n",
       "      <th>atm_withdrawal</th>\n",
       "      <th>...</th>\n",
       "      <th>nationality_134</th>\n",
       "      <th>nationality_141</th>\n",
       "      <th>nationality_172</th>\n",
       "      <th>nationality_176</th>\n",
       "      <th>nationality_191</th>\n",
       "      <th>is_pep_0</th>\n",
       "      <th>is_pep_1</th>\n",
       "      <th>category_0</th>\n",
       "      <th>category_1</th>\n",
       "      <th>category_2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer</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",
       "      <th></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",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>90000000</th>\n",
       "      <td>734958.58</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>79.0</td>\n",
       "      <td>24.63</td>\n",
       "      <td>124.0</td>\n",
       "      <td>3</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.128703</td>\n",
       "      <td>215959.85</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000001</th>\n",
       "      <td>1009400.35</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>78.49</td>\n",
       "      <td>79.0</td>\n",
       "      <td>3</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.635584</td>\n",
       "      <td>53.97</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000002</th>\n",
       "      <td>11051.68</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>77.0</td>\n",
       "      <td>35.74</td>\n",
       "      <td>72.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.138152</td>\n",
       "      <td>1632.91</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000003</th>\n",
       "      <td>31193.29</td>\n",
       "      <td>81.0</td>\n",
       "      <td>0.308642</td>\n",
       "      <td>77.0</td>\n",
       "      <td>2.05</td>\n",
       "      <td>95.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.333904</td>\n",
       "      <td>3203.10</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90000004</th>\n",
       "      <td>205489.77</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>39.0</td>\n",
       "      <td>147.53</td>\n",
       "      <td>148.0</td>\n",
       "      <td>4</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.341578</td>\n",
       "      <td>5626.65</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 32 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            turnover  transaction_count  io_ratio   age  \\\n",
       "customer                                                  \n",
       "90000000   734958.58                7.0  0.142857  79.0   \n",
       "90000001  1009400.35                2.0  0.500000   0.0   \n",
       "90000002    11051.68                3.0  0.666667  77.0   \n",
       "90000003    31193.29               81.0  0.308642  77.0   \n",
       "90000004   205489.77                1.0  1.000000  39.0   \n",
       "\n",
       "          inactive_days_average  inactive_days_max  n_of_accounts  \\\n",
       "customer                                                            \n",
       "90000000                  24.63              124.0              3   \n",
       "90000001                  78.49               79.0              3   \n",
       "90000002                  35.74               72.0              2   \n",
       "90000003                   2.05               95.0              2   \n",
       "90000004                 147.53              148.0              4   \n",
       "\n",
       "          distinct_counterparties  channel_risk  atm_withdrawal     ...      \\\n",
       "customer                                                            ...       \n",
       "90000000                      6.0      0.128703       215959.85     ...       \n",
       "90000001                      1.0      0.635584           53.97     ...       \n",
       "90000002                      1.0      0.138152         1632.91     ...       \n",
       "90000003                      2.0      0.333904         3203.10     ...       \n",
       "90000004                      1.0      0.341578         5626.65     ...       \n",
       "\n",
       "          nationality_134  nationality_141  nationality_172  nationality_176  \\\n",
       "customer                                                                       \n",
       "90000000                1                0                0                0   \n",
       "90000001                0                0                0                0   \n",
       "90000002                0                0                0                0   \n",
       "90000003                0                0                0                0   \n",
       "90000004                0                0                1                0   \n",
       "\n",
       "          nationality_191  is_pep_0  is_pep_1  category_0  category_1  \\\n",
       "customer                                                                \n",
       "90000000                0         1         0           1           0   \n",
       "90000001                0         1         0           0           1   \n",
       "90000002                1         1         0           1           0   \n",
       "90000003                0         1         0           1           0   \n",
       "90000004                0         1         0           1           0   \n",
       "\n",
       "          category_2  \n",
       "customer              \n",
       "90000000           0  \n",
       "90000001           0  \n",
       "90000002           0  \n",
       "90000003           0  \n",
       "90000004           0  \n",
       "\n",
       "[5 rows x 32 columns]"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "The least populated class in y has only 1 member, which is too few. The minimum number of groups for any class cannot be less than 2.",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-87-a06f17f97f7a>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[0mx_res_index\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mx\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mreset_index\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      2\u001b[0m \u001b[0my_res_index\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mreset_index\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 3\u001b[1;33m \u001b[1;32mfor\u001b[0m \u001b[0ma\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mb\u001b[0m \u001b[1;32min\u001b[0m \u001b[0msss\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msplit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mx_res_index\u001b[0m\u001b[1;33m,\u001b[0m\u001b[0my_res_index\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      4\u001b[0m     \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0my_res_index\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0ma\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msuspicious\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtolist\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcount\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'suspicious'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      5\u001b[0m     \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0my_res_index\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0ma\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msuspicious\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtolist\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcount\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'non suspicious'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py\u001b[0m in \u001b[0;36msplit\u001b[1;34m(self, X, y, groups)\u001b[0m\n\u001b[0;32m   1235\u001b[0m         \"\"\"\n\u001b[0;32m   1236\u001b[0m         \u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mgroups\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mindexable\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mgroups\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1237\u001b[1;33m         \u001b[1;32mfor\u001b[0m \u001b[0mtrain\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtest\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_iter_indices\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mgroups\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1238\u001b[0m             \u001b[1;32myield\u001b[0m \u001b[0mtrain\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtest\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1239\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py\u001b[0m in \u001b[0;36m_iter_indices\u001b[1;34m(self, X, y, groups)\u001b[0m\n\u001b[0;32m   1583\u001b[0m         \u001b[0mclass_counts\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbincount\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0my_indices\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1584\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmin\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mclass_counts\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m<\u001b[0m \u001b[1;36m2\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1585\u001b[1;33m             raise ValueError(\"The least populated class in y has only 1\"\n\u001b[0m\u001b[0;32m   1586\u001b[0m                              \u001b[1;34m\" member, which is too few. The minimum\"\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1587\u001b[0m                              \u001b[1;34m\" number of groups for any class cannot\"\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mValueError\u001b[0m: The least populated class in y has only 1 member, which is too few. The minimum number of groups for any class cannot be less than 2."
     ]
    }
   ],
   "source": [
    "x_res_index = x.reset_index()\n",
    "y_res_index = y.reset_index()\n",
    "for a, b in sss.split(x_res_index,y_res_index):\n",
    "    print(y_res_index[a].suspicious.tolist().count('suspicious'))\n",
    "    print(y_res_index[a].suspicious.tolist().count('non suspicious'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline_logisticcv = Pipeline([\n",
    "    ('classifier',          LogisticRegressionCV(Cs=9,class_weight='balanced',cv=sss))\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "     steps=[('classifier', LogisticRegressionCV(Cs=9, class_weight='balanced',\n",
       "           cv=StratifiedShuffleSplit(n_splits=5, random_state=0, test_size=0.2,\n",
       "            train_size=None),\n",
       "           dual=False, fit_intercept=True, intercept_scaling=1.0,\n",
       "           max_iter=100, multi_class='warn', n_jobs=None, penalty='l2',\n",
       "           random_state=None, refit=True, scoring=None, solver='lbfgs',\n",
       "           tol=0.0001, verbose=0))])"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipeline_logisticcv.fit(train_x,train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                precision    recall  f1-score   support\n",
      "\n",
      "non suspicious       0.99      0.03      0.05    198157\n",
      "    suspicious       0.01      0.98      0.02      1843\n",
      "\n",
      "     micro avg       0.04      0.04      0.04    200000\n",
      "     macro avg       0.50      0.50      0.04    200000\n",
      "  weighted avg       0.98      0.04      0.05    200000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "evaluate_classifier(pipeline_logisticcv, test_x,test_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline_gboosting = Pipeline([\n",
    "    ('classifier',          GradientBoostingClassifier())\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Pipeline(memory=None,\n",
       "     steps=[('classifier', GradientBoostingClassifier(criterion='friedman_mse', init=None,\n",
       "              learning_rate=0.1, loss='deviance', max_depth=3,\n",
       "              max_features=None, max_leaf_nodes=None,\n",
       "              min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "              min_samples_leaf=...    subsample=1.0, tol=0.0001, validation_fraction=0.1,\n",
       "              verbose=0, warm_start=False))])"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipeline_gboosting.fit(train_x,train_y)\n",
    "pipeline_gboosting.fit(test_x,test_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                precision    recall  f1-score   support\n",
      "\n",
      "non suspicious       0.99      1.00      1.00    198157\n",
      "    suspicious       0.88      0.09      0.16      1843\n",
      "\n",
      "     micro avg       0.99      0.99      0.99    200000\n",
      "     macro avg       0.94      0.54      0.58    200000\n",
      "  weighted avg       0.99      0.99      0.99    200000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate_classifier(pipeline_gboosting, test_x,test_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 275,
   "metadata": {},
   "outputs": [],
   "source": [
    "predicted_labels = gb_smote.predict_proba(data_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 276,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_set_with_label = data_test.copy(deep=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 277,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_set_with_label['label'] = predicted_labels[:,1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_set_with_label = data_set_with_label.sort_values(by='label',ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "metadata": {},
   "outputs": [],
   "source": [
    "aaa = data_set_with_label.head(1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 280,
   "metadata": {},
   "outputs": [],
   "source": [
    "aaa.index.to_frame().to_csv(\"res.csv\",index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1000000"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(predicted_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "23627"
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(predicted_labels[predicted_labels == 'suspicious'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [],
   "source": [
    "parameters = {\n",
    "    'loss': ('deviance', 'exponential'),\n",
    "    'n_estimators' : (100,200),\n",
    "    'max_depth':(3,4,5)\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [],
   "source": [
    "grid_search = GridSearchCV(pipeline_gboosting, parameters, cv=5, n_jobs=-1, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#pipeline_gboosting.get_params()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#grid_search.fit(train_x, train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from scipy.stats import randint as sp_randint\n",
    "from sklearn.model_selection import RandomizedSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'X' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-145-b77c45afc311>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      9\u001b[0m \u001b[0mplt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfigure\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfigsize\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m8\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m6\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     10\u001b[0m \u001b[0mplt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtitle\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"Feature importances\"\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 11\u001b[1;33m \u001b[0mplt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbar\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mimportances\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mindices\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcolor\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;34m\"b\"\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0malign\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;34m\"center\"\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     12\u001b[0m \u001b[0mplt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mxticks\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mrotation\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m90\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     13\u001b[0m \u001b[0mplt\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mxticks\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mX\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mcolumns\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mindices\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'X' is not defined"
     ]
    },
    {
     "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": [
    "# build a random forest and compute the feature importances\n",
    "clf_forest = RandomForestClassifier(n_estimators=30, random_state=0)\n",
    "clf_forest.fit(train_x, train_y)\n",
    "importances = clf_forest.feature_importances_\n",
    "indices = np.argsort(importances)[::-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "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 the feature importances of the forest\n",
    "plt.figure(figsize=(8, 6))\n",
    "plt.title(\"Feature importances\")\n",
    "plt.bar(range(train_x.shape[1]), importances[indices], color=\"b\", align=\"center\")\n",
    "plt.xticks(rotation=90)\n",
    "plt.xticks(range(train_x.shape[1]), train_x.columns[indices])\n",
    "plt.xlim([-1, train_x.shape[1]])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                precision    recall  f1-score   support\n",
      "\n",
      "non suspicious       0.99      1.00      1.00    198157\n",
      "    suspicious       0.80      0.04      0.07      1843\n",
      "\n",
      "     micro avg       0.99      0.99      0.99    200000\n",
      "     macro avg       0.90      0.52      0.53    200000\n",
      "  weighted avg       0.99      0.99      0.99    200000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate_classifier(clf_forest, test_x,test_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-149-43bd3c1f0ed5>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m     10\u001b[0m                                    n_iter=n_iter_search, cv=5)\n\u001b[0;32m     11\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 12\u001b[1;33m \u001b[0mrandom_search\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mtrain_x\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtrain_y\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_search.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, X, y, groups, **fit_params)\u001b[0m\n\u001b[0;32m    720\u001b[0m                 \u001b[1;32mreturn\u001b[0m \u001b[0mresults_container\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    721\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 722\u001b[1;33m             \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_run_search\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mevaluate_candidates\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    723\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    724\u001b[0m         \u001b[0mresults\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mresults_container\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_search.py\u001b[0m in \u001b[0;36m_run_search\u001b[1;34m(self, evaluate_candidates)\u001b[0m\n\u001b[0;32m   1513\u001b[0m         evaluate_candidates(ParameterSampler(\n\u001b[0;32m   1514\u001b[0m             \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mparam_distributions\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mn_iter\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1515\u001b[1;33m             random_state=self.random_state))\n\u001b[0m",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_search.py\u001b[0m in \u001b[0;36mevaluate_candidates\u001b[1;34m(candidate_params)\u001b[0m\n\u001b[0;32m    709\u001b[0m                                \u001b[1;32mfor\u001b[0m \u001b[0mparameters\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0mtrain\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtest\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    710\u001b[0m                                in product(candidate_params,\n\u001b[1;32m--> 711\u001b[1;33m                                           cv.split(X, y, groups)))\n\u001b[0m\u001b[0;32m    712\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    713\u001b[0m                 \u001b[0mall_candidate_params\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mextend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mcandidate_params\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py\u001b[0m in \u001b[0;36m__call__\u001b[1;34m(self, iterable)\u001b[0m\n\u001b[0;32m    984\u001b[0m                 \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_iterating\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_original_iterator\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    985\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 986\u001b[1;33m             \u001b[1;32mwhile\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdispatch_one_batch\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0miterator\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    987\u001b[0m                 \u001b[1;32mpass\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    988\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py\u001b[0m in \u001b[0;36mdispatch_one_batch\u001b[1;34m(self, iterator)\u001b[0m\n\u001b[0;32m    823\u001b[0m                 \u001b[1;32mreturn\u001b[0m \u001b[1;32mFalse\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    824\u001b[0m             \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 825\u001b[1;33m                 \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_dispatch\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mtasks\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    826\u001b[0m                 \u001b[1;32mreturn\u001b[0m \u001b[1;32mTrue\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    827\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py\u001b[0m in \u001b[0;36m_dispatch\u001b[1;34m(self, batch)\u001b[0m\n\u001b[0;32m    780\u001b[0m         \u001b[1;32mwith\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_lock\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    781\u001b[0m             \u001b[0mjob_idx\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_jobs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 782\u001b[1;33m             \u001b[0mjob\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_backend\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mapply_async\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mbatch\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcallback\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcb\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    783\u001b[0m             \u001b[1;31m# A job can complete so quickly than its callback is\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    784\u001b[0m             \u001b[1;31m# called before we get here, causing self._jobs to\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\_parallel_backends.py\u001b[0m in \u001b[0;36mapply_async\u001b[1;34m(self, func, callback)\u001b[0m\n\u001b[0;32m    180\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mapply_async\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfunc\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcallback\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    181\u001b[0m         \u001b[1;34m\"\"\"Schedule a func to be run\"\"\"\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 182\u001b[1;33m         \u001b[0mresult\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mImmediateResult\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfunc\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    183\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mcallback\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    184\u001b[0m             \u001b[0mcallback\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mresult\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\_parallel_backends.py\u001b[0m in \u001b[0;36m__init__\u001b[1;34m(self, batch)\u001b[0m\n\u001b[0;32m    543\u001b[0m         \u001b[1;31m# Don't delay the application, to avoid keeping the input\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    544\u001b[0m         \u001b[1;31m# arguments in memory\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 545\u001b[1;33m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mresults\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mbatch\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    546\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    547\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mget\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py\u001b[0m in \u001b[0;36m__call__\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    259\u001b[0m         \u001b[1;32mwith\u001b[0m \u001b[0mparallel_backend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_backend\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    260\u001b[0m             return [func(*args, **kwargs)\n\u001b[1;32m--> 261\u001b[1;33m                     for func, args, kwargs in self.items]\n\u001b[0m\u001b[0;32m    262\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    263\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m__len__\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py\u001b[0m in \u001b[0;36m<listcomp>\u001b[1;34m(.0)\u001b[0m\n\u001b[0;32m    259\u001b[0m         \u001b[1;32mwith\u001b[0m \u001b[0mparallel_backend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_backend\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    260\u001b[0m             return [func(*args, **kwargs)\n\u001b[1;32m--> 261\u001b[1;33m                     for func, args, kwargs in self.items]\n\u001b[0m\u001b[0;32m    262\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    263\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m__len__\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_validation.py\u001b[0m in \u001b[0;36m_fit_and_score\u001b[1;34m(estimator, X, y, scorer, train, test, verbose, parameters, fit_params, return_train_score, return_parameters, return_n_test_samples, return_times, return_estimator, error_score)\u001b[0m\n\u001b[0;32m    526\u001b[0m             \u001b[0mestimator\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX_train\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m**\u001b[0m\u001b[0mfit_params\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    527\u001b[0m         \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 528\u001b[1;33m             \u001b[0mestimator\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX_train\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my_train\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m**\u001b[0m\u001b[0mfit_params\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    529\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    530\u001b[0m     \u001b[1;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[1;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\ensemble\\forest.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, X, y, sample_weight)\u001b[0m\n\u001b[0;32m    333\u001b[0m                     \u001b[0mt\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0msample_weight\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mi\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mtrees\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    334\u001b[0m                     verbose=self.verbose, class_weight=self.class_weight)\n\u001b[1;32m--> 335\u001b[1;33m                 for i, t in enumerate(trees))\n\u001b[0m\u001b[0;32m    336\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    337\u001b[0m             \u001b[1;31m# Collect newly grown trees\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py\u001b[0m in \u001b[0;36m__call__\u001b[1;34m(self, iterable)\u001b[0m\n\u001b[0;32m    984\u001b[0m                 \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_iterating\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_original_iterator\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    985\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 986\u001b[1;33m             \u001b[1;32mwhile\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdispatch_one_batch\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0miterator\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    987\u001b[0m                 \u001b[1;32mpass\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    988\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py\u001b[0m in \u001b[0;36mdispatch_one_batch\u001b[1;34m(self, iterator)\u001b[0m\n\u001b[0;32m    823\u001b[0m                 \u001b[1;32mreturn\u001b[0m \u001b[1;32mFalse\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    824\u001b[0m             \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 825\u001b[1;33m                 \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_dispatch\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mtasks\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    826\u001b[0m                 \u001b[1;32mreturn\u001b[0m \u001b[1;32mTrue\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    827\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py\u001b[0m in \u001b[0;36m_dispatch\u001b[1;34m(self, batch)\u001b[0m\n\u001b[0;32m    780\u001b[0m         \u001b[1;32mwith\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_lock\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    781\u001b[0m             \u001b[0mjob_idx\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_jobs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 782\u001b[1;33m             \u001b[0mjob\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_backend\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mapply_async\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mbatch\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcallback\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcb\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    783\u001b[0m             \u001b[1;31m# A job can complete so quickly than its callback is\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    784\u001b[0m             \u001b[1;31m# called before we get here, causing self._jobs to\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\_parallel_backends.py\u001b[0m in \u001b[0;36mapply_async\u001b[1;34m(self, func, callback)\u001b[0m\n\u001b[0;32m    180\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mapply_async\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfunc\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcallback\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    181\u001b[0m         \u001b[1;34m\"\"\"Schedule a func to be run\"\"\"\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 182\u001b[1;33m         \u001b[0mresult\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mImmediateResult\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mfunc\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    183\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mcallback\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    184\u001b[0m             \u001b[0mcallback\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mresult\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\_parallel_backends.py\u001b[0m in \u001b[0;36m__init__\u001b[1;34m(self, batch)\u001b[0m\n\u001b[0;32m    543\u001b[0m         \u001b[1;31m# Don't delay the application, to avoid keeping the input\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    544\u001b[0m         \u001b[1;31m# arguments in memory\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 545\u001b[1;33m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mresults\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mbatch\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    546\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    547\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mget\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py\u001b[0m in \u001b[0;36m__call__\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    259\u001b[0m         \u001b[1;32mwith\u001b[0m \u001b[0mparallel_backend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_backend\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    260\u001b[0m             return [func(*args, **kwargs)\n\u001b[1;32m--> 261\u001b[1;33m                     for func, args, kwargs in self.items]\n\u001b[0m\u001b[0;32m    262\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    263\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m__len__\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\externals\\joblib\\parallel.py\u001b[0m in \u001b[0;36m<listcomp>\u001b[1;34m(.0)\u001b[0m\n\u001b[0;32m    259\u001b[0m         \u001b[1;32mwith\u001b[0m \u001b[0mparallel_backend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_backend\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    260\u001b[0m             return [func(*args, **kwargs)\n\u001b[1;32m--> 261\u001b[1;33m                     for func, args, kwargs in self.items]\n\u001b[0m\u001b[0;32m    262\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    263\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m__len__\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\ensemble\\forest.py\u001b[0m in \u001b[0;36m_parallel_build_trees\u001b[1;34m(tree, forest, X, y, sample_weight, tree_idx, n_trees, verbose, class_weight)\u001b[0m\n\u001b[0;32m    119\u001b[0m             \u001b[0mcurr_sample_weight\u001b[0m \u001b[1;33m*=\u001b[0m \u001b[0mcompute_sample_weight\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m'balanced'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mindices\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    120\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 121\u001b[1;33m         \u001b[0mtree\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0msample_weight\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcurr_sample_weight\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcheck_input\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mFalse\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    122\u001b[0m     \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    123\u001b[0m         \u001b[0mtree\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0msample_weight\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0msample_weight\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcheck_input\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mFalse\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\tree\\tree.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, X, y, sample_weight, check_input, X_idx_sorted)\u001b[0m\n\u001b[0;32m    797\u001b[0m             \u001b[0msample_weight\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0msample_weight\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    798\u001b[0m             \u001b[0mcheck_input\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcheck_input\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 799\u001b[1;33m             X_idx_sorted=X_idx_sorted)\n\u001b[0m\u001b[0;32m    800\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    801\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\Anaconda3\\lib\\site-packages\\sklearn\\tree\\tree.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, X, y, sample_weight, check_input, X_idx_sorted)\u001b[0m\n\u001b[0;32m    363\u001b[0m                                            min_impurity_split)\n\u001b[0;32m    364\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 365\u001b[1;33m         \u001b[0mbuilder\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbuild\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtree_\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0msample_weight\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mX_idx_sorted\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    366\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    367\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mn_outputs_\u001b[0m \u001b[1;33m==\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "clf = RandomForestClassifier(n_estimators=20)\n",
    "param_dist = {\"max_depth\": [3, None],\n",
    "              \"max_features\": sp_randint(1, 11),\n",
    "              \"min_samples_split\": sp_randint(2, 11),\n",
    "              \"bootstrap\": [True, False],\n",
    "              \"criterion\": [\"gini\", \"entropy\"]}\n",
    "\n",
    "n_iter_search = 20\n",
    "random_search = RandomizedSearchCV(clf, param_distributions=param_dist,\n",
    "                                   n_iter=n_iter_search, cv=5)\n",
    "\n",
    "random_search.fit(train_x, train_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_over, test_over = train_test_split(data,test_size=0.2, random_state= 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Class count\n",
    "count_class_0, count_class_1 = train_over.suspicious.value_counts()\n",
    "\n",
    "# Divide by class\n",
    "df_class_0 = train_over[train_over['suspicious'] == 'non suspicious']\n",
    "df_class_1 = train_over[train_over['suspicious'] == 'suspicious']\n",
    "\n",
    "df_class_1_over = df_class_1.sample(10000, replace=True)\n",
    "df_test_over = pd.concat([df_class_0, df_class_1_over], axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_frequencies = []\n",
    "label_frequencies.append(df_test_over.suspicious.tolist().count('suspicious')/len(df_test_over))\n",
    "label_frequencies.append(df_test_over.suspicious.tolist().count('non suspicious')/len(df_test_over))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.012457085341000254, 0.9875429146589998]"
      ]
     },
     "execution_count": 184,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_frequencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "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>suspicious</th>\n",
       "      <th>turnover</th>\n",
       "      <th>transaction_count</th>\n",
       "      <th>io_ratio</th>\n",
       "      <th>age</th>\n",
       "      <th>inactive_days_average</th>\n",
       "      <th>inactive_days_max</th>\n",
       "      <th>n_of_accounts</th>\n",
       "      <th>distinct_counterparties</th>\n",
       "      <th>channel_risk</th>\n",
       "      <th>...</th>\n",
       "      <th>nationality_134</th>\n",
       "      <th>nationality_141</th>\n",
       "      <th>nationality_172</th>\n",
       "      <th>nationality_176</th>\n",
       "      <th>nationality_191</th>\n",
       "      <th>is_pep_0</th>\n",
       "      <th>is_pep_1</th>\n",
       "      <th>category_0</th>\n",
       "      <th>category_1</th>\n",
       "      <th>category_2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer</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",
       "      <th></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",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>90771718</th>\n",
       "      <td>non suspicious</td>\n",
       "      <td>409868.95</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>60.0</td>\n",
       "      <td>30.46</td>\n",
       "      <td>92.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.410764</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90521462</th>\n",
       "      <td>non suspicious</td>\n",
       "      <td>143538.46</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>26.0</td>\n",
       "      <td>94.21</td>\n",
       "      <td>189.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.184895</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90137361</th>\n",
       "      <td>non suspicious</td>\n",
       "      <td>770447.14</td>\n",
       "      <td>26.0</td>\n",
       "      <td>0.269231</td>\n",
       "      <td>48.0</td>\n",
       "      <td>6.11</td>\n",
       "      <td>98.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.130168</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90404985</th>\n",
       "      <td>non suspicious</td>\n",
       "      <td>2407.86</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>18.0</td>\n",
       "      <td>49.89</td>\n",
       "      <td>150.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.216123</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90910092</th>\n",
       "      <td>non suspicious</td>\n",
       "      <td>1093644.44</td>\n",
       "      <td>18.0</td>\n",
       "      <td>0.555556</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7.87</td>\n",
       "      <td>79.0</td>\n",
       "      <td>5</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.607511</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 33 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              suspicious    turnover  transaction_count  io_ratio   age  \\\n",
       "customer                                                                  \n",
       "90771718  non suspicious   409868.95                3.0  0.666667  60.0   \n",
       "90521462  non suspicious   143538.46                3.0  0.000000  26.0   \n",
       "90137361  non suspicious   770447.14               26.0  0.269231  48.0   \n",
       "90404985  non suspicious     2407.86                4.0  0.250000  18.0   \n",
       "90910092  non suspicious  1093644.44               18.0  0.555556   0.0   \n",
       "\n",
       "          inactive_days_average  inactive_days_max  n_of_accounts  \\\n",
       "customer                                                            \n",
       "90771718                  30.46               92.0              1   \n",
       "90521462                  94.21              189.0              1   \n",
       "90137361                   6.11               98.0              1   \n",
       "90404985                  49.89              150.0              1   \n",
       "90910092                   7.87               79.0              5   \n",
       "\n",
       "          distinct_counterparties  channel_risk     ...      nationality_134  \\\n",
       "customer                                            ...                        \n",
       "90771718                      1.0      0.410764     ...                    0   \n",
       "90521462                      1.0      0.184895     ...                    0   \n",
       "90137361                      1.0      0.130168     ...                    0   \n",
       "90404985                      1.0      0.216123     ...                    0   \n",
       "90910092                      1.0      0.607511     ...                    0   \n",
       "\n",
       "          nationality_141  nationality_172  nationality_176  nationality_191  \\\n",
       "customer                                                                       \n",
       "90771718                0                1                0                0   \n",
       "90521462                0                1                0                0   \n",
       "90137361                0                0                0                0   \n",
       "90404985                0                0                0                0   \n",
       "90910092                0                1                0                0   \n",
       "\n",
       "          is_pep_0  is_pep_1  category_0  category_1  category_2  \n",
       "customer                                                          \n",
       "90771718         1         0           1           0           0  \n",
       "90521462         1         0           1           0           0  \n",
       "90137361         1         0           1           0           0  \n",
       "90404985         1         0           1           0           0  \n",
       "90910092         1         0           0           1           0  \n",
       "\n",
       "[5 rows x 33 columns]"
      ]
     },
     "execution_count": 185,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_test_over.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_over = df_test_over.suspicious\n",
    "x_over = df_test_over.drop('suspicious',axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingClassifier(criterion='friedman_mse', init=None,\n",
       "              learning_rate=0.1, loss='deviance', max_depth=3,\n",
       "              max_features=None, max_leaf_nodes=None,\n",
       "              min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "              min_samples_leaf=1, min_samples_split=2,\n",
       "              min_weight_fraction_leaf=0.0, n_estimators=100,\n",
       "              n_iter_no_change=None, presort='auto', random_state=None,\n",
       "              subsample=1.0, tol=0.0001, validation_fraction=0.1,\n",
       "              verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 187,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gb =  GradientBoostingClassifier()\n",
    "gb.fit(x_over,y_over)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_over_test = test_over.suspicious\n",
    "x_over_test = test_over.drop('suspicious',axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                precision    recall  f1-score   support\n",
      "\n",
      "non suspicious       0.99      1.00      1.00    198157\n",
      "    suspicious       0.66      0.06      0.11      1843\n",
      "\n",
      "     micro avg       0.99      0.99      0.99    200000\n",
      "     macro avg       0.82      0.53      0.55    200000\n",
      "  weighted avg       0.99      0.99      0.99    200000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAocAAAHeCAYAAADpfaFAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3Xm8XfO5+PHP4yQxF0VVRY2hRcklVJUaS4zRVltuTS1VWq22lxrqUsOvaJVOVCnXeMXQIm0pNQ8tScxijKBiVkNFQiTn+f2x1tndcrOSI84+++Ssz/u+9qt7f9faaz8nuTvn8Tzf73dFZiJJkiQBzNPuACRJktR3mBxKkiSpweRQkiRJDSaHkiRJajA5lCRJUoPJoSRJkhpMDiVJktRgcihJkqQGk0NJkiQ1DGh3AJq5AYOW8dY1UhvME9HuEKRamvr2xF798r3z8oQe+z07cIkV+9U/HFYOJUmS1GDlUJIk1U/n9HZH0GeZHEqSpPrJznZH0GfZVpYkSVKDlUNJklQ/nVYOq5gcSpKk2knbypVsK0uSJKnByqEkSaof28qVTA4lSVL92FauZFtZkiRJDVYOJUlS/bgJdiWTQ0mSVD+2lSvZVpYkSVKDlUNJklQ/rlauZHIoSZJqx02wq9lWliRJUoOVQ0mSVD+2lSuZHEqSpPqxrVzJtrIkSZIarBxKkqT6cRPsSiaHkiSpfmwrV7KtLEmSpAYrh5IkqX5crVzJ5FCSJNWPbeVKtpUlSZLUYOVQkiTVj23lSiaHkiSpdjLdyqaKbWVJkiQ1WDmUJEn144KUSiaHkiSpfpxzWMm2siRJkhqsHEqSpPqxrVzJ5FCSJNVPp6uVq9hWliRJUoOVQ0mSVD+2lSuZHEqSpPpxtXIl28qSJElqsHIoSZLqx7ZyJZNDSZJUP7aVK9lWliRJUoOVQ0mSVD9WDiuZHEqSpNrJdBPsKraVJUmS1GDlUJIk1Y9t5Uomh5IkqX7cyqaSbWVJkiQ1WDmUJEn1Y1u5ksmhJEmqH9vKlWwrS5IkqcHKoSRJqh/bypVMDiVJUv3YVq5kcihJkurHymEl5xxKkiSpwcqhJEmqHyuHlUwOJUlS/TjnsJJtZUmSJDVYOZQkSfVjW7mSyaEkSaof28qVbCtLkiSpwcqhJEmqH9vKlUwOJUlS/dhWrmRbWZIkSQ1WDiVJUv3YVq5kcihJkurH5LCSbWVJkiQ1WDmUJEn1k9nuCPosk0NJklQ/tpUr2VaWJElSg5VDSZJUP1YOK1k5lCRJ9ZOdPffohog4KyJejIgHZhj/dkQ8EhHjIuInTeOHRsT48thWTePDy7HxEXFI0/gKEXFHRDwWERdFxKByfN7y9fjy+PKzi9XkUJIkqfXOBoY3D0TEpsAIYM3MXB04sRxfDdgZWL18z6kR0RERHcApwNbAasAu5bkAJwAnZ+YQ4FVgr3J8L+DVzFwZOLk8b5ZMDiVJUv10dvbcoxsy82bglRmG9wOOz8y3y3NeLMdHACMz8+3MfAIYD6xXPsZn5oTMnAqMBEZERACbAZeW7z8H2LHpWueUzy8FNi/Pr2RyKEmS6iezxx4RsU9EjG167NPNKFYBNirbvTdFxLrl+DLA003nTSzHqsYXB17LzGkzjL/rWuXx18vzK7kgRZIk6X3IzNOB0+fgrQOAxYD1gXWBiyNiRWBmlb1k5kW9nMX5zOZYZVCSJEn10jdWK08E/pCZCYyOiE5giXJ82abzBgPPls9nNv4ysGhEDCirg83nd11rYkQMABbh/7a338W2siRJqp9ennNY4XKKuYJExCrAIIpEbxSwc7nSeAVgCDAaGAMMKVcmD6JYtDKqTC5vAHYqr7sHcEX5fFT5mvL49eX5lawcSpIktVhEXAhsAiwREROBI4GzgLPK7W2mAnuUidu4iLgYeBCYBnwrM6eX19kfuBroAM7KzHHlRxwMjIyIY4G7gTPL8TOB8yJiPEXFcOfZxjqb5FFtMmDQMv7FSG0wz6wX8UlqkalvT+zVL9+U332/x37Pzr/3Sf3qHw4rh5IkqXay0xpMFeccSpIkqcHKoSRJqp++sVq5TzI5lCRJ9dPNeyLXkW1lSZIkNVg5lCRJ9eOClEomh5IkqX6cc1jJtrIkSZIarBxKkqT6sXJYyeRQkiTVj3eIq2RbWZIkSQ1WDiVJUv3YVq5k5VB6j844/Wc8O/Fe7rn7usbYF76wHffecz1T33qaddZeszH+wQ8uxrXXXMJrrzzKL35+bDvCleZqp//2RCY+fQ9333VtY+yC809lzOirGTP6ah595O+MGX01AMOGDW2Mjx1zDSN2GA7A4MFLc83VF3PfvTdwz93Xsf/+e7XlZ1Ef05k99+hnTA6bRMSVEbHoLI7vGxG792ZM6nvOPfditt3uK+8aGzfuYb74pa9zyy23v2v8rbfe4sgf/YQfHHxMb4Yo9RvnnncJ222/67vGvrLrN1l3va1Yd72tuOzyK7n88quA4nu4/qe2Yd31tmK77XfllFOOp6Ojg2nTpvODg49mzbU2ZcONdmC/fffg4x8b0o4fR5or2FZukpnbzOb4ab0Vi/quW269g+WWG/yusYcfHj/TcydPnsJtfxvDSiut0BuhSf3OrTP5vjXb6Qvbs9XwLwMwZcpbjfH55puXLBccPP/8izz//IsATJr0Jg8//BgfWebDPPTwYy2MXH2et8+rNNdVDiNiwYj4c0TcGxEPRMSXI+LJiFiiPD4sIm4sn28cEfeUj7sjYuGI2CQibo6IyyLiwYg4LSLmKc9vvs7uEXFf+TnnlWM/iogDy+dDI+L28pzLImKxcvzGiBhWPl8iIp4sn68eEaPLWO6LCP+zVZLehw03/CQvvvgS48c/0Rhbd93/4J67r+OuO69l//0PZfr06e96z3LLDWattdZg9Oi7eztc9TW2lSvNdckhMBx4NjPXysw1gL/M4twDgW9l5lBgI2BKOb4e8F/AJ4CVgM83vykiVgd+CGyWmWsBB8zk2ucCB2fmmsD9wJGziXtf4BdlLMOAiTOeEBH7RMTYiBjb2fnmbC4nSfX25S+P4KKLr3jX2JgxdzP0PzZng09vyw9+sD/zzjtv49iCCy7ARSNP58ADf8Qbb0zq7XClucbcmBzeD2wRESdExEaZ+foszr0NOCkivgMsmpnTyvHRmTkhM6cDFwIbzvC+zYBLM/NlgMx8pflgRCxSXu+mcugc4DOzifvvwGERcTCwXGZOmfGEzDw9M4dl5rB55llwNpeTpPrq6OhgxxFbc8klf5zp8YcfHs+bb05m9dVXBWDAgAFcdNHpXDjyMi6/4qreDFV9VHZ29tijv5nrksPMfBRYhyJJPC4ijgCm8e+fZb6mc48H9gbmB26PiI91HZrxsjO8jpmMdVdVLP8L7EBRvbw6Ijabw+tLUu1tvvlGPPLI4zzzzHONseWXX5aOjg4APvrRZVhllRV56qmngWLV88MPj+cXvzijLfGqD7KtXGmuSw4j4iPA5Mw8HzgRWBt4kiJhBPhC07krZeb9mXkCMBboSg7Xi4gVyrmGXwZuneFjrgO+FBGLl9f5YPPBslr5akRsVA7tBnRVEZtj2akplhWBCZn5S2AU8O/9TjRXOf+8U7j15lGsuspKPDlhLF/dc2dGjBjOkxPGsv766zDqinO58k8XNM4f/+jtnPjTI9hj9y/x5ISxfPzjTjeVuuu8c3/NzTddwSqrrMSEx8ew5547A/ClL+7ARRdf/q5zP73Betw59hrGjL6aSy7+Hd854If885+vssEG67Lrrjux6Safbmx1M3y4/30uVYmcy24fExFbAT8FOoF3gP0oKoNnAi8AdwDDMnOTiPgVsCkwHXgQ2BP4FHAE8BLFnMObgW9mZme5eGRYZr4cEXsAB5XvvTsz94yIHwGTMvPEiBgKnAYsAEwAvpqZr5bVyYuBScD1wK6ZuXxEHArsWsb8PPCfM7armw0YtMzc9Rcj9RPzRLQ7BKmWpr49sVe/fG8eu2uP/Z5d8PDz+9U/HHNdcvh+RcQmwIGZuV27Y5kVk0OpPUwOpfbo9eTw6K/0XHJ4xAX96h+Oua6tLEmSpNap3SbYmXkjcGObw5AkSe3UD1cZ95TaJYeSJEn9cZVxT7GtLEmSpAYrh5IkqX68t3Ilk0NJklQ/tpUrmRxKkqTa6Y+3vespzjmUJElSg5VDSZJUP7aVK5kcSpKk+jE5rGRbWZIkSQ1WDiVJUv24lU0lk0NJklQ/tpUr2VaWJElSg5VDSZJUO2nlsJLJoSRJqh+Tw0q2lSVJktRg5VCSJNWPt8+rZHIoSZLqx7ZyJdvKkiRJarByKEmS6sfKYSWTQ0mSVDuZJodVbCtLkiSpwcqhJEmqH9vKlUwOJUlS/ZgcVrKtLEmSpAYrh5IkqXa8t3I1k0NJklQ/JoeVbCtLkiSpwcqhJEmqH2+tXMnkUJIk1Y5zDqvZVpYkSVKDlUNJklQ/Vg4rmRxKkqT6cc5hJdvKkiRJarByKEmSascFKdVMDiVJUv3YVq5kW1mSJEkNVg4lSVLt2FauZnIoSZLqx7ZyJdvKkiRJarByKEmSaietHFYyOZQkSfVjcljJtrIkSZIarBxKkqTasa1czeRQkiTVj8lhJdvKkiRJLRYRZ0XEixHxQNPYTyPi4Yi4LyIui4hFm44dGhHjI+KRiNiqaXx4OTY+Ig5pGl8hIu6IiMci4qKIGFSOz1u+Hl8eX352sZocSpKk2snOnnt009nA8BnG/gqskZlrAo8ChwJExGrAzsDq5XtOjYiOiOgATgG2BlYDdinPBTgBODkzhwCvAnuV43sBr2bmysDJ5XmzZHIoSZJqp7eTw8y8GXhlhrFrMnNa+fJ2YHD5fAQwMjPfzswngPHAeuVjfGZOyMypwEhgREQEsBlwafn+c4Adm651Tvn8UmDz8vxKJoeSJEnvQ0TsExFjmx77zMFlvgZcVT5fBni66djEcqxqfHHgtaZEs2v8Xdcqj79enl/JBSmSJKl2enK1cmaeDpw+p++PiB8C04ALuoZm9jHMvKiXszh/VteqZHIoSZLqJ2fZWe01EbEHsB2weWZ2JW0TgWWbThsMPFs+n9n4y8CiETGgrA42n991rYkRMQBYhBna2zOyrSxJktQGETEcOBjYITMnNx0aBexcrjReARgCjAbGAEPKlcmDKBatjCqTyhuAncr37wFc0XStPcrnOwHXNyWhM2XlUJIk1U5vb4IdERcCmwBLRMRE4EiK1cnzAn8t14jcnpn7Zua4iLgYeJCi3fytzJxeXmd/4GqgAzgrM8eVH3EwMDIijgXuBs4sx88EzouI8RQVw51nG+tskke1yYBBy/gXI7XBPLNexCepRaa+PbFXv3zPbbhpj/2eXfrWG/rVPxy2lSVJktRgW1mSJNWO91auZnIoSZJqJ/vIauW+yLayJEmSGqwcSpKk2rGtXM3kUJIk1U522lauYnIoSZJqx538qjnnUJIkSQ1WDiVJUu3YVq5mcihJkmrH5LCabWVJkiQ1VFYOI+IDs3pjZv6r58ORJElqPRekVJtVW3kckEBz3bXrdQIfbWFckiRJLWNbuVplcpiZy/ZmIJIkSWq/bs05jIidI+Kw8vngiFintWFJkiS1Tmb02KO/mW1yGBG/BjYFdiuHJgOntTIoSZKkVsrOnnv0N93ZymaDzFw7Iu4GyMxXImJQi+OSJElSG3QnOXwnIuahWIRCRCwO9MM8WZIk1UVnP2wH95TuJIenAL8HloyIo4AvAUe1NCpJkqQW6o9zBXvKbJPDzDw3Iu4EtiiHvpiZD7Q2LEmSJLVDd2+f1wG8Q9Fa9q4qkiRpruY+h9W6s1r5h8CFwEeAwcD/RsShrQ5MkiSpVTJ77tHfdKdyuCuwTmZOBoiI/wfcCRzXysAkSZLU+7qTHD41w3kDgAmtCUeSJKn1bCtXq0wOI+JkijmGk4FxEXF1+XpL4NbeCU+SJKnnuZVNtVlVDrtWJI8D/tw0fnvrwpEkSVI7VSaHmXlmbwYiSZLUW9znsNps5xxGxErA/wNWA+brGs/MVVoYlyRJUsv0x1XGPaU7exaeDfwPEMDWwMXAyBbGJEmSpDbpTnK4QGZeDZCZj2fm4cCmrQ1LkiSpdTozeuzR33RnK5u3IyKAxyNiX+AZ4EOtDUuSJKl1nHNYrTvJ4feAhYDvUMw9XAT4WiuDkiRJUnvMNjnMzDvKp28Au7U2HEmSpNZzQUq1WW2CfRnFptczlZmfb0lEkiRJLdYf5wr2lFlVDn/da1FIUh/x5jM3tzsESWqrWW2CfV1vBiJJktRbXJBSrTsLUiRJkvoV28rVurPPoSRJkmqi25XDiJg3M99uZTCSJEm9wcXK1WZbOYyI9SLifuCx8vVaEfGrlkcmSZLUIt4hpVp32sq/BLYD/gmQmffi7fMkSZL6pe60lefJzKeKO+g1TG9RPJIkSS3nauVq3UkOn46I9YCMiA7g28CjrQ1LkiSpdTrbHUAf1p228n7A94GPAi8A65djkiRJ6me6c2/lF4GdeyEWSZKkXpHYVq4y2+QwIs5gJiu+M3OflkQkSZLUYp3uZVOpO3MOr216Ph/wOeDp1oQjSZKkdupOW/mi5tcRcR7w15ZFJEmS1GKdtpUrzcm9lVcAluvpQCRJknqLcw6rdWfO4av8e87hPMArwCGtDEqSJEntMcvkMIqdr9cCnimHOjPTKZySJGmu5j6H1Wa5z2GZCF6WmdPLh4mhJEma6yXRY4/+pjubYI+OiLVbHokkSZLarrKtHBEDMnMasCHw9Yh4HHgTCIqiogmjJEmaK9lWrjarOYejgbWBHXspFkmSpF5hclhtVslhAGTm470UiyRJktpsVsnhkhHx/aqDmXlSC+KRJElquf64kKSnzCo57AAWAv/0JElS/9JpdlNpVsnhc5l5dK9FIkmSpLab7ZxDSZKk/sZ7K1ebVXK4ea9FIUmS1Iu8q0e1yuQwM1/pzUAkSZJ6i1vZVOvOHVIkSZJUE7NqK0uSJPVLneGcwyomh5IkqXacc1jNtrIkSZIarBxKkqTacUFKNZNDSZJUO94hpZptZUmSpBaLiO9FxLiIeCAiLoyI+SJihYi4IyIei4iLImJQee685evx5fHlm65zaDn+SERs1TQ+vBwbHxGHvJ9YTQ4lSVLtdBI99pidiFgG+A4wLDPXADqAnYETgJMzcwjwKrBX+Za9gFczc2Xg5PI8ImK18n2rA8OBUyOiIyI6gFOArYHVgF3Kc+eIyaEkSaqd7MFHNw0A5o+IAcACwHPAZsCl5fFzgB3L5yPK15THN4+IKMdHZubbmfkEMB5Yr3yMz8wJmTkVGFmeO0dMDiVJkt6HiNgnIsY2PfZpPp6ZzwAnAv+gSApfB+4EXsvMaeVpE4FlyufLAE+X751Wnr948/gM76kanyMuSJEkSbXTkwtSMvN04PSq4xGxGEUlbwXgNeASihbw/7lU11sqjlWNz6zYN8dbOZocSpKk2unlrWy2AJ7IzJcAIuIPwAbAohExoKwODgaeLc+fCCwLTCzb0IsArzSNd2l+T9X4e2ZbWZIkqbX+AawfEQuUcwc3Bx4EbgB2Ks/ZA7iifD6qfE15/PrMzHJ853I18wrAEGA0MAYYUq5+HkSxaGXUnAZr5VCSJNVOb94+LzPviIhLgbuAacDdFG3oPwMjI+LYcuzM8i1nAudFxHiKiuHO5XXGRcTFFInlNOBbmTkdICL2B66mWAl9VmaOm9N4o0hE1dcMGLSMfzFSG0x59pZ2hyDV0sAlVuzVbanPHLxrj/2e3Wvi+f1qS23bypIkSWqwrSxJkmrHeytXMzmUJEm1Y3JYzbayJEmSGqwcSpKk2sl+tYSkZ5kcSpKk2rGtXM22siRJkhqsHEqSpNqxcljN5FCSJNWOd5qoZltZkiRJDVYOJUlS7XS6WrmSyaEkSaod5xxWs60sSZKkBiuHkiSpdqwcVjM5lCRJteNq5Wq2lSVJktRg5VCSJNWOq5WrmRxKkqTacc5hNdvKkiRJarByKEmSascFKdVMDiVJUu10mh5Wsq0sSZKkBiuHkiSpdlyQUs3kUJIk1Y5N5Wq2lSVJktRg5VCSJNWObeVqJoeSJKl2vENKNdvKkiRJarByKEmSasd9DquZHEqSpNoxNaxmW1mSJEkNVg4lSVLtuFq5msmhJEmqHeccVrOtLEmSpAYrh5IkqXasG1YzOZQkSbXjnMNqtpUlSZLUYOVQkiTVjgtSqpkcSpKk2jE1rGZyKEmSasc5h9WccyhJkqQGK4eSJKl20sZyJZNDSZJUO7aVq9lWliRJUoOVQ0mSVDtuZVPN5FCSJNWOqWE128qSJElqsHIoSZJqx7ZyNZND6X0a/+jtvDFpEtOndzJt2jTW/9Q2rLXW6pz66+OZd755mTZtGt/+9mGMGXtPu0OV+qzDf3wSN982mg8utiiXn38aAA8/NoFjfvorJk95i48s/SFOOPIHLLTggrwzbRpHHvdzHnr0caZNn84Owzfn67t/GYBbbx/L8T8/jemdnXxh++HsvduXALjjzns48de/4513prHaqitz9KHfY8CADl7/1xv893En8/QzzzHvoEEcc9j3GLLi8u36Y1AvcrVyNdvKTSLib7M5fnREbNFb8WjuscVnv8iwdbdk/U9tA8DxP/4hxxx7EsPW3ZKjjjqR44/7YZsjlPq2Hbf5LKeddOy7xo48/ud8d7+vctl5v2Hzz2zA/1zwewCuuf4Wpr7zDped9xsuPuuXXHLFlTzz3AtMnz6dY392Cr/52TGMuuC3XHntjTz+xFN0dnZy2LE/46dHHcLl55/GRz78Ia646loAzjj3Ij42ZCUuO/c3/Pi/D+T4n5/W6z+71NeYHDbJzA1mc/yIzLy2t+LR3CszWfgDCwPwgUUW5tnnXmhzRFLfNmzoJ1ik/M50efIfExk29BMAfGrdtfnrTbcCEBFMeestpk2bzttvT2XgwIEstOAC3P/Qo3x08EdYdpmlGThwIFtvvjHX33I7r73+LwYNHMjyHx3cuNa1NxbXevzJf7D+OmsBsOJyy/LMcy/w8iuv9taPrTbKHvy//qYlyWFELB8RD0XEGRExLiKuiYj5y2NDI+L2iLgvIi6LiMXK8Rsj4oSIGB0Rj0bERjO57tIRcXNE3BMRD3SdExGTms7ZKSLOLp9/sTzv3oi4uRzbMyKuiIi/RMQjEXFk03ubr/ODiLi/fO/x5djZEbFT+XzziLi7POesiJi3HH8yIpYonw+LiBvL5xuXcd9Tvu/d/wpqrpWZXHXlhdxx+1XsvddXAPj+gUdywnGH88TjY/jJ8f/NDw8/rs1RSnOflVdcnhtuvR2Aa264hedfeBmAz266IfPPNx+bjvhPPvv53dlzl8+zyAcW5sWXXubDH1qy8f6lPrQEL770TxZbdBGmTZvOAw89Wlzrxlt5/sXiWquuvCLX3lQ0je5/8BGee+FFXiiPqX/r7MFHf9PKyuEQ4JTMXB14DfhCOX4ucHBmrgncDxzZ9J4Bmbke8N0Zxrv8J3B1Zg4F1gJmN4nrCGCrzFwL2KFpfD3gK8BQ4IsRMaz5TRGxNbAj8MnyvT+Z4fh8wNnAlzPzExRzN/ebTSwHAt8qY98ImDLjCRGxT0SMjYixnZ1vzuZy6is+s8mOrPfJ4Wy3/a7st9+ebLThJ/nGPrvzXwf9iBVWWpf/Ougozvjtz9odpjTXOeaw73Hh7//Il772bd6cPIWBA4tp8vc/+Agd88zD9VdcwF8uPZtzLvwDTz/zHDmTAk5EUWn86dGH8JNfns7Oex/AggvMT0dH8etv792+yL/emMQX9vgWF1w6io8NWYmOjo7e/DGlPqeVC1KeyMyu5O1OYPmIWARYNDNvKsfPAS5pes8fms+fyTXHAGdFxEDg8qbrV7kNODsiLm66NsBfM/OfABHxB2BDYGzT8S2A/8nMyQCZ+coM1121/Pkebfo5vgX8fDaxnBQRFwB/yMyJM56QmacDpwMMGLRM/6tT91PPlS3jl176J1dccRXrrjuU3Xf7It/7/hEAXHrpHzn9tJ+2M0RprrTicstyxs9/DBQt5pv/NhqAK/96I59efxgDBwxg8cUWZeiaqzHu4cf48FJL8vyLLzXe/8KLL7PkEosDMHSNj3Pub04E4LY77uSpp58BYKEFF+TYH34fKLoAW+20J4M/slSv/Yxqn/7YDu4prawcvt30fDrdS0S73jPT8zPzZuAzwDPAeRGxe9ehptPmazp/X+BwYFngnohYfCbnz+x1zGRsxuNVpvHvP9fmWI4H9gbmB26PiI/N4hqaSyywwPwstNCCjeef3WJjxo17hGefe4GNP/MpADbbdEMeG/9EO8OU5kr/fPU1ADo7O/ntOSP50o7Fgq+ll1qS0XfeS2Yyecpb3DfuYVZYblnW+Ngq/GPis0x89nneeecdrrruJjbdcP13XWvq1KmcdcEljWv9641JvPPOOwD8/o9/YZ2hn2ChBRfs7R9VbWBbuVqvbmWTma9HxKsRsVFm3gLsBtw0u/d1iYjlgGcy84yIWBBYm6JN/UJEfBx4BPgc8EZ5/kqZeQdwR0RsT5EkAnw2Ij5I0drdEfjaDB91DXBERPxvZk6OiA/OUD18mKISunJmjp/h53gSWAe4in+30rtiuR+4PyI+BXysvI7mYksttSSXXnImAAMGdDBy5OVcfc2NTNr3IE466WgGDBjA22+9xX77/aDNkUp920FHHs+Yu+/jtdf+xeY77so399qNyVOmMPIPfwJgi4034HPbbgnALp/fnsN/fBI77rovSbLjNluy6sorAHDY9/bjG98/nOnTp/O57bZk5RWXA+B/LriUm/42muzs5Muf25ZPrjMUgAlPPc1hx5xIxzzzsOLyH+XoQ7/bhp9e6lsiZzZJ4/1eNGJ54E+ZuUb5+kBgocz8UUQMBU4DFgAmAF/NzFfLhRsHZubYckHH2Mxcfobr7gEcBLwDTAJ2z8wnykUiJwBPAw+Un7Vn2TIeQlHpu45iLuMewDbAgsDKwP9m5lHl9Sdl5kLl80OA3YGpwJWZeVi50OVPmXlpRGwOnEiRYI8B9svMt8tFMmcCLwB3AMMyc5OI+BWwKUVV9EFgz8xsrq6+i21dqHrHAAAU6UlEQVRlqT2mPHtLu0OQamngEivOqivX43Zb7vM99nv2vKf+0Kuxt1pLksO+LCL2pEjY9m93LLNicii1h8mh1B69nRzu2oPJ4fn9LDl0n0NJkiQ11O72eZl5NsU2NJIkqaa8t3K12iWHkiRJbmVTzbayJEmSGqwcSpKk2umP+xP2FJNDSZJUO845rGZbWZIkSQ1WDiVJUu24IKWayaEkSaod5xxWs60sSZLUCyKiIyLujog/la9XiIg7IuKxiLgoIgaV4/OWr8eXx5dvusah5fgjEbFV0/jwcmx8eQvgOWZyKEmSaicze+zxHhwAPNT0+gTg5MwcArwK7FWO7wW8mpkrAyeX5xERqwE7A6sDw4FTy4SzAzgF2BpYDdilPHeOmBxKkqTa6SR77NEdETEY2Bb4Xfk6gM2AS8tTzgF2LJ+PKF9THt+8PH8EMDIz387MJ4DxwHrlY3xmTsjMqcDI8tw5YnIoSZLUej8HfsC/pzsuDryWmdPK1xOBZcrnywBPA5THXy/Pb4zP8J6q8TlicihJkmqnswcfEbFPRIxteuzT/FkRsR3wYmbe2Tw8k7ByNsfe6/gccbWyJEmqnZ7cyiYzTwdOn8UpnwZ2iIhtgPmAD1BUEheNiAFldXAw8Gx5/kRgWWBiRAwAFgFeaRrv0vyeqvH3zMqhJElSC2XmoZk5ODOXp1hQcn1mfgW4AdipPG0P4Iry+ajyNeXx67NY+TIK2LlczbwCMAQYDYwBhpSrnweVnzFqTuO1cihJkmqnj9w+72BgZEQcC9wNnFmOnwmcFxHjKSqGOwNk5riIuBh4EJgGfCszpwNExP7A1UAHcFZmjpvToOI9LsFWLxkwaBn/YqQ2mPLsLe0OQaqlgUusOLN5cy2z9bJb99jv2auevqpXY28128qSJElqsK0sSZJqx9vnVTM5lCRJtdOTq5X7G9vKkiRJarByKEmSaqePrFbuk0wOJUlS7bhbSzXbypIkSWqwcihJkmrHtnI1k0NJklQ7rlauZltZkiRJDVYOJUlS7XS6IKWSyaEkSaodU8NqtpUlSZLUYOVQkiTVjquVq5kcSpKk2jE5rGZbWZIkSQ1WDiVJUu14+7xqJoeSJKl2bCtXMzmUJEm14x1SqjnnUJIkSQ1WDiVJUu0457CayaEkSaod5xxWs60sSZKkBiuHkiSpdmwrVzM5lCRJtWNbuZptZUmSJDVYOZQkSbXjPofVTA4lSVLtdDrnsJJtZUmSJDVYOZQkSbVjW7mayaEkSaod28rVbCtLkiSpwcqhJEmqHdvK1UwOJUlS7dhWrmZbWZIkSQ1WDiVJUu3YVq5mcihJkmrHtnI128qSJElqsHIoSZJqx7ZyNZNDSZJUO5md7Q6hz7KtLEmSpAYrh5IkqXY6bStXMjmUJEm1k65WrmRbWZIkSQ1WDiVJUu3YVq5mcihJkmrHtnI128qSJElqsHIoSZJqx9vnVTM5lCRJteMdUqrZVpYkSVKDlUNJklQ7LkipZnIoSZJqx61sqtlWliRJUoOVQ0mSVDu2lauZHEqSpNpxK5tqtpUlSZLUYOVQkiTVjm3laiaHkiSpdlytXM22siRJkhqsHEqSpNqxrVzN5FCSJNWOq5Wr2VaWJElSg5VDSZJUO+mClEomh5IkqXZsK1ezrSxJkqQGK4eSJKl2XK1czeRQkiTVjnMOq9lWliRJUoPJoSRJqp3M7LFHd0TE8Ih4JCLGR8QhLf7x3hfbypIkqXZ6c85hRHQApwCfBSYCYyJiVGY+2GtBvAdWDiVJklprPWB8Zk7IzKnASGBEm2OqZOWwj5o29ZlodwyacxGxT2ae3u44pLrxu6fueqd3f88uAzzd9Hoi8Mle/Pz3xMqh1Br7tDsAqab87qnXRcQ+ETG26THj/x/OLBHts8ulrRxKkiS9D2W1elYV64nAsk2vBwPPtjSo98HKoSRJUmuNAYZExAoRMQjYGRjV5pgqWTmUWsM5T1J7+N1Tn5OZ0yJif+BqoAM4KzPHtTmsSuHtYyRJktTFtrIkSZIaTA4lSZLUYHIoSZKkBpNDSZIkNZgcSn1ARHhHHKkP6vpu+h1VnZgcSm0WEZGZGRHDI+KLETGw3TFJetd3c3vgF343VRcmh1Kblb98tgFOAv6Zme+0OyZJje/mtsBRwB/9bqouTA6lNoqIeSJiYeD7wHcy8/qI2Li8T+c67Y5PEsOBQ4F7I2K7iDg3IjaLiEG2mtVfeYcUqb0WzszXI+Je4AvlDvqvAx8GPgTc2dbopJqKiCGZ+RjwEvB1YEngWoqiyq7ATeldJNRPeYcUqU0iYjBwCzACeBPYHHggM/9WtrK+DXwBmOwvIan3RMRCwAXA3RQt5Y2BlzJzXER8DDgT+HJmTmxjmFLL2FaW2qT8xfJzYCSwaGaeXiaGmwInAL/MzDdNDKXWm6FFPIUiKVwTOCIzbywTwx2B3wPHmxiqP7NyKPWyiFgVeCMzny1ffwM4CPgScA9FwviXzLyyfVFK9RMRGwJvZebYiOgAVqNIEu/KzGPLaR+PZOZf2xqo1GImh1IvKX/ZBHAxxTymo4DnyhWRvwB2Aj4LPJ6Zb7cvUqmeIuJA4HvADpl5Z7l1ze7AIcAZmfmTtgYo9RLbylKLNbWr5s3MacBuwCCKXzhLl8duBEYDS5oYSr2jaYPrpSNiQGaeSPG9vDAihpVb1zwJjAJual+kUu+ycii1UNMmup8F9geeAB6kmND+O2Aq8DSwA/C1zHyg6z1tC1qqgabv5jYU0zpeBiYCPwa2AA4DrqSY7rFrZt7WtmClXmZyKLVARHRk5vTy+frAORS/bKYB+wJjKNrKXwVWBG7LzD+3KVypNiJiUGZOLZ+vTJEA7kpRzV+HYmXyV4B1gaWAlzPzhjaFK7WF+xxKPSwilgI2i4hLy7bUosDIzPx9efzvwNnAmpn5u6b3WTGUWigilga2i4gLMnMyxe/AezNzdHn8MWB1YPPM/FMbQ5XayjmHUs/7GDAW+EBELAm8BewZER8ByMwXgWeARZrfZGIotVwHxd6iC5f7FT4FrBwRBwBk5gsUm9APaV+IUvuZHEo9pGtye2beRDF36RiKtvFNwK+AayJi7Yj4NDAMmNyuWKW6iYh5yr0JHwH+H8VdTz4EfIui0n9yRGwGbE0x7UOqLeccSj2kaYL7asDDwPbApsAEigUoewNbUlQvTs3MUW0LVqqRpu/mMOA1irm/h1BU8EdR3KHoRxRbTN3gd1N1Z3Io9aBy5eORwNcz876IGEFxW7xHgd9m5jsRsUBmTnaOodR7ImJ74AjgwMy8KSJWokgQnwb+JzOfbjrX76ZqzeRQ6iERMQS4HNgrM29vGt+BYquaccCpwFR/8UitNcOOAUsDFwLfz8y7uo5FxIoUuwb8AzguMye1MWSpz3C1sjSHyl8swzPz1HJocWB8V2LYtWVGZo6KiMkUd0Nxg2upxcqFYOdGxIhy25qBFHPs/9l1Svm//6DY43AJE0Pp31yQIs25N4A7y61roGgdz1u2r8jMqRGxRUQclpnXZua4tkUq1UhmvkSx0GSliBicmf8A7gM2iIgPZua08j7KvwEmZeYD7YxX6mtsK0tzoFz52BkRA4B7gWsy83sRsT+wPDCF4pZ4vwS+l5nXtC1YqUa6vpvl858A+1FsNL8GxebWA4G7gO8A+2fmVe2KVeqrTA6l96hp5WPXvKVlgMuAKyi2rFmXYgub14CrM/OPbQxXqo2m7+bCmflGOXYE8DVgbWAhYBuKKSB/884n0syZHEpzICK2AHaiuE/yVRT3Zb0WuCIzjy7PGViuTnblo9RLImJbYC9gOvC7zLw6Ig6n+A+2z2TmM34npVlzQYr0HpX3Sv4RcDHwcWBNipWQWwK3R8RCmfmD8tZ53vlE6iXlPob/RfH93BjYKiJWzsxjI+KDwF0R8VGKfQ6nty9SqW+zcii9B+UK5f8FzszMMyLiw8AmFBWJb0bEYGCFzLylnXFKdRMRywInAVMyc/dybATFnVD2yswXImLVzHyknXFKcwNXK0uz0XVbvPJ/pwPPAd+KiMUy83ngBmD1iPh4Zk7MzFu63iOpdZq+mx+guLvJWOATZVJIZl4BdAKfLt/yWDvilOY2JofSLDRNcN8B+CnF7bYOAG4GTi0XoywAfBBolOFtJUut1fTd3Ab4MzAIOAf4HfC5iNg/Ij4OrAI8BdC1ilnSrDnnUJqF8pfPthRzmA4t90d7mqJ9dQwwBrgF2C8zH25fpFK9lN/NzYGfAftm5r+Af0XE+RTb1RxAcevKvTLzzuYtbiTNmsmhNHvbUNyD9e5yg+sRwPnA0RT3ZV02M28F78kq9bLVKe6X/HhE7AF8GzgLuAR4h2LB2Hxg1VB6L2wrSxUiYuXy6cvA3hS/cLr2SvtPYAJwBjA5Ik4t9z00MZRaLCJWj4gVKO5KdBRFO3kx4NfAl4EOiu/rE8D2EbFQu2KV5kauVpZmIiIWpliVfDtwHLAp8EJmPlDOYzoT+DzwAsUdUd7MzBfbFK5UG2Wi91/AEOCbwKLA1Mx8vpwDfDmwa2Y+EhEfoVi9/Gr7IpbmPlYOpdIMK4wnU1Qk1gUOy8zrysRwBHApcFxmPp+FJ0wMpdZp/m5m5iSKBPBRirm/85WJ4c4UG9L/uGu7msx81sRQeu+sHEpNImIDYFJm3hcRHRT3Yz0KuD0zj4+IA4AHM/Ovzi+Uek9EbATskpnfLF+vAXwOWIZi/u9gYKHMvN7vpvT+uCBFtde0JcYKwJ7AlhExIjPvjYgHgSuBgyJiamae1PU+f/lIrTVDknc/cGn5PfxuWclfjmKB2DHANzJzGvjdlN4v28qqvaZ9DH8P/DfF1hgXRsQa5S3wJgCjgL+1MUypVpr3MYyIgzLzNWBVigUmvy5Pexa4GzipKzGU9P7ZVlbtRcRQ4GyKltVD5dh5wEcpNrvelWKC+21tC1KqoYjYiuI/1r6ZmTeXYx8A/g48BHwS2Dszr25flFL/Y3Ko2itXHx9M8QtnKYp7JT8LTKNoKb+cmde3LUCphiJiIHAhxZY0FwPbAusDf6X4rq5LMT/43rYFKfVTJoeqvXJrjD2BXSiqFI8CnwFezcwL2xiaVGsR8R1gY2ARikrhAOCdzPxOWwOT+jmTQ6kUEYMyc2pEDKNoMx+Qmde1OSyptiJiJWBFij1G74uIjSn2Hd0+M//Z3uik/svkUCqVW9cMBU6l2CvtijaHJNVSebeh6TOMbUZxB5SDMvPP7YlMqgeTQ6lJRCwIfCgzn3CvNKn1mlYlfxxIYGJmToqIAV0rkCNiCYqFYQ+5+ERqPZNDSVJblauSzwauBZYDdsrMF5sriBExX2a+1cYwpdpwn0NJUttExGrA1sAXMnM3ipXIf4mIJTNzekQMADAxlHqPyaEkqddFREe5U8AvKPYrnASQmQcD1wC3RsSH3Nxa6n0mh5KkXhMR0fU0MycBewMvAdtGxCIAmXkI8CdglfZEKdWbcw4lSb2iafHJFsDuwFPArcCdwPnA9cAZmflqG8OUas/KoSSpV5SJ4cbAr4AbgH8AxwDDgT2AHYBvdM0zlNQeVg4lSb0mIr4CLJWZJ5Wv16S4M9EuwIeBBTJzdBtDlGrPyqEkqWWa5hh2mZ9iz8IuD1Dcy3zpzHzAxFBqP5NDSVLLdLWSI+KbEbFaZv4OuCMirouIDwLDgDXx95HUZ9hWliT1uKbFJ58EzgIeBCZTLEC5APgJsDywOHBcZo5qV6yS3s3kUJLUEhGxHnA08IPMvC8idgHWB+7LzDMjYh5g0cx8xdtVSn2HZXxJUqssCmwBfLZ8fQlwG7B+RBwABPAqFO3ntkQo6f9wuwBJUktk5jUR8XnguIh4NjMvjIhLgQ7g3q77JkvqW0wOJUktk5mjImIacExEDMrMc4AL2x2XpGrOOZQktVxE7AAcT9Fmfj4zO9sckqQKJoeSpF4REUtm5kvtjkPSrJkcSpIkqcHVypIkSWowOZQkSVKDyaEkSZIaTA4lSZLUYHIoqa0iYnpE3BMRD0TEJRGxwPu41iYR8afy+Q4Rccgszl00Ir45B5/xo4g4sLvjM5xzdkTs9B4+a/mIeOC9xihJ74fJoaR2m5KZQzNzDWAqsG/zwSi853+rMnNUZh4/i1MWBd5zcihJ/Z3JoaS+5BZg5bJi9lBEnArcBSwbEVtGxN8j4q6ywrgQQEQMj4iHI+JW4PNdF4qIPSPi1+XzpSLisoi4t3xsQLEh80pl1fKn5XkHRcSYiLgvIo5qutYPI+KRiLgWWHV2P0REfL28zr0R8fsZqqFbRMQtEfFoRGxXnt8RET9t+uxvvN8/SEmaUyaHkvqEiBgAbA3cXw6tCpybmf8BvAkcDmyRmWsDY4HvR8R8wBnA9sBGwIcrLv9L4KbMXAtYGxgHHAI8XlYtD4qILYEhwHrAUGCdiPhMRKwD7Az8B0XyuW43fpw/ZOa65ec9BOzVdGx5YGNgW+C08mfYC3g9M9ctr//1iFihG58jST3OeytLarf5I+Ke8vktwJnAR4CnMvP2cnx9YDXgtogAGAT8HfgY8ERmPgYQEecD+8zkMzYDdgfIzOnA6xGx2AznbFk+7i5fL0SRLC4MXJaZk8vPGNWNn2mNiDiWonW9EHB107GLy1vHPRYRE8qfYUtgzab5iIuUn/1oNz5LknqUyaGkdpuSmUObB8oE8M3mIeCvmbnLDOcNBXrqNk8BHJeZv53hM747B59xNrBjZt4bEXsCmzQdm/FaWX72tzOzOYkkIpZ/j58rSe+bbWVJc4PbgU9HxMoAEbFARKwCPAysEBErleftUvH+64D9yvd2RMQHgDcoqoJdrga+1jSXcZmI+BBwM/C5iJg/IhamaGHPzsLAcxExEPjKDMe+GBHzlDGvCDxSfvZ+5flExCoRsWA3PkeSepyVQ0l9Xma+VFbgLoyIecvhwzPz0YjYB/hzRLwM3AqsMZNLHACcHhF7AdOB/TLz7xFxW7lVzFXlvMOPA38vK5eTgF0z866IuAi4B3iKovU9O/8N3FGefz/vTkIfAW4ClgL2zcy3IuJ3FHMR74riw18Cduzen44k9azI7KmOjCRJkuZ2tpUlSZLUYHIoSZKkBpNDSZIkNZgcSpIkqcHkUJIkSQ0mh5IkSWowOZQkSVLD/weuSr2BB3wA4QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate_classifier(gb, x_over_test,y_over_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [],
   "source": [
    "from imblearn.over_sampling import SMOTE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_smote, test_smote = train_test_split(data,test_size=0.2, random_state= 1)\n",
    "x_train_smote = train_smote.drop('suspicious',axis=1)\n",
    "y_train_smote = train_smote.suspicious"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 270,
   "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>suspicious</th>\n",
       "      <th>turnover</th>\n",
       "      <th>transaction_count</th>\n",
       "      <th>io_ratio</th>\n",
       "      <th>age</th>\n",
       "      <th>inactive_days_average</th>\n",
       "      <th>inactive_days_max</th>\n",
       "      <th>n_of_accounts</th>\n",
       "      <th>distinct_counterparties</th>\n",
       "      <th>channel_risk</th>\n",
       "      <th>...</th>\n",
       "      <th>nationality_134</th>\n",
       "      <th>nationality_141</th>\n",
       "      <th>nationality_172</th>\n",
       "      <th>nationality_176</th>\n",
       "      <th>nationality_191</th>\n",
       "      <th>is_pep_0</th>\n",
       "      <th>is_pep_1</th>\n",
       "      <th>category_0</th>\n",
       "      <th>category_1</th>\n",
       "      <th>category_2</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>customer</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",
       "      <th></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",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>0 rows × 33 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [suspicious, turnover, transaction_count, io_ratio, age, inactive_days_average, inactive_days_max, n_of_accounts, distinct_counterparties, channel_risk, atm_withdrawal, atm_deposit, nationality_18, nationality_32, nationality_33, nationality_45, nationality_47, nationality_50, nationality_66, nationality_94, nationality_113, nationality_117, nationality_123, nationality_134, nationality_141, nationality_172, nationality_176, nationality_191, is_pep_0, is_pep_1, category_0, category_1, category_2]\n",
       "Index: []\n",
       "\n",
       "[0 rows x 33 columns]"
      ]
     },
     "execution_count": 270,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_smote.loc[train_smote['turnover'] == 649096.4932921625]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "metadata": {},
   "outputs": [],
   "source": [
    "smote = SMOTE(ratio='minority',random_state=10)\n",
    "X_sm, y_sm = smote.fit_sample(x_train_smote, y_train_smote)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "metadata": {},
   "outputs": [],
   "source": [
    "c = np.c_[X_sm.reshape(len(X_sm), -1), y_sm.reshape(len(y_sm), -1)]\n",
    "np.random.shuffle(c)\n",
    "X_sm_2 = c[:, :X_sm.size//len(X_sm)].reshape(X_sm.shape)\n",
    "y_sm_2 = c[:, X_sm.size//len(X_sm):].reshape(y_sm.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_sm_2[:100]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 271,
   "metadata": {},
   "outputs": [],
   "source": [
    "gb_smote =  GradientBoostingClassifier()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 272,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingClassifier(criterion='friedman_mse', init=None,\n",
       "              learning_rate=0.1, loss='deviance', max_depth=3,\n",
       "              max_features=None, max_leaf_nodes=None,\n",
       "              min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "              min_samples_leaf=1, min_samples_split=2,\n",
       "              min_weight_fraction_leaf=0.0, n_estimators=100,\n",
       "              n_iter_no_change=None, presort='auto', random_state=None,\n",
       "              subsample=1.0, tol=0.0001, validation_fraction=0.1,\n",
       "              verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 272,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gb_smote.fit(X_sm_2,y_sm_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 273,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_smote_test = test_smote.suspicious\n",
    "x_smote_test = test_smote.drop('suspicious',axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 274,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                precision    recall  f1-score   support\n",
      "\n",
      "non suspicious       0.99      1.00      0.99    198157\n",
      "    suspicious       0.22      0.10      0.14      1843\n",
      "\n",
      "     micro avg       0.99      0.99      0.99    200000\n",
      "     macro avg       0.60      0.55      0.57    200000\n",
      "  weighted avg       0.98      0.99      0.99    200000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate_classifier(gb_smote, x_smote_test,y_smote_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1.43538460e+05, 3.00000000e+00, 0.00000000e+00, 2.60000000e+01,\n",
       "       9.42100000e+01, 1.89000000e+02, 1.00000000e+00, 1.00000000e+00,\n",
       "       1.84894851e-01, 0.00000000e+00, 1.21257302e+03, 0.00000000e+00,\n",
       "       0.00000000e+00, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "       0.00000000e+00, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "       0.00000000e+00, 0.00000000e+00, 0.00000000e+00, 0.00000000e+00,\n",
       "       1.00000000e+00, 0.00000000e+00, 0.00000000e+00, 1.00000000e+00,\n",
       "       0.00000000e+00, 1.00000000e+00, 0.00000000e+00, 0.00000000e+00])"
      ]
     },
     "execution_count": 247,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_sm[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "792756"
      ]
     },
     "execution_count": 234,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(y_sm[y_sm == 'suspicious'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1585512"
      ]
     },
     "execution_count": 230,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(y_sm == 'suspicious')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_poly, test_poly = train_test_split(data,test_size=0.2, random_state= 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train_poly = train_poly.drop('suspicious',axis=1)\n",
    "y_train_poly = train_poly.suspicious"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train_poly['poly_turno'] = x_train_poly['turnover']**2\n",
    "x_train_poly['poly_inactive_days_average'] = x_train_poly['inactive_days_average']**2\n",
    "x_train_poly['poly_channel_risk'] = x_train_poly['channel_risk']**2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingClassifier(criterion='friedman_mse', init=None,\n",
       "              learning_rate=0.1, loss='deviance', max_depth=3,\n",
       "              max_features=None, max_leaf_nodes=None,\n",
       "              min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "              min_samples_leaf=1, min_samples_split=2,\n",
       "              min_weight_fraction_leaf=0.0, n_estimators=100,\n",
       "              n_iter_no_change=None, presort='auto', random_state=None,\n",
       "              subsample=1.0, tol=0.0001, validation_fraction=0.1,\n",
       "              verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 217,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gb_poly =  GradientBoostingClassifier()\n",
    "gb_poly.fit(x_train_poly,y_train_poly)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_test_poly = test_poly.drop('suspicious',axis=1)\n",
    "y_test_poly = test_poly.suspicious"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_test_poly['poly_turno'] = x_test_poly['turnover']**2\n",
    "x_test_poly['poly_inactive_days_average'] = x_test_poly['inactive_days_average']**2\n",
    "x_test_poly['poly_channel_risk'] = x_test_poly['channel_risk']**2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                precision    recall  f1-score   support\n",
      "\n",
      "non suspicious       0.99      1.00      1.00    198157\n",
      "    suspicious       0.61      0.05      0.09      1843\n",
      "\n",
      "     micro avg       0.99      0.99      0.99    200000\n",
      "     macro avg       0.80      0.52      0.54    200000\n",
      "  weighted avg       0.99      0.99      0.99    200000\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "evaluate_classifier(gb_poly, x_test_poly,y_test_poly)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  },
  "toc": {
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}