fraud_detection_ann.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "--6vF7fEX19s"
   },
   "source": [
    "### Fraud Detection & Anomaly Detection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "executionInfo": {
     "elapsed": 1191,
     "status": "ok",
     "timestamp": 1734378872564,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "YiVPOnUFCT67"
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {
    "executionInfo": {
     "elapsed": 5222,
     "status": "ok",
     "timestamp": 1734376301939,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "pidvWNBc4WyT"
   },
   "outputs": [],
   "source": [
    "data= pd.read_csv(\"data/fraudTrain.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 878
    },
    "executionInfo": {
     "elapsed": 583,
     "status": "ok",
     "timestamp": 1734376305179,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "4F6p-E6_4b2x",
    "outputId": "04a722bb-deef-4e26-b6a3-1d3ba7fe03a2"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>trans_date_trans_time</th>\n",
       "      <th>cc_num</th>\n",
       "      <th>merchant</th>\n",
       "      <th>category</th>\n",
       "      <th>amt</th>\n",
       "      <th>first</th>\n",
       "      <th>last</th>\n",
       "      <th>gender</th>\n",
       "      <th>street</th>\n",
       "      <th>...</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "      <th>city_pop</th>\n",
       "      <th>job</th>\n",
       "      <th>dob</th>\n",
       "      <th>trans_num</th>\n",
       "      <th>unix_time</th>\n",
       "      <th>merch_lat</th>\n",
       "      <th>merch_long</th>\n",
       "      <th>is_fraud</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>2019-01-01 00:00:18</td>\n",
       "      <td>2703186189652095</td>\n",
       "      <td>fraud_Rippin, Kub and Mann</td>\n",
       "      <td>misc_net</td>\n",
       "      <td>4.97</td>\n",
       "      <td>Jennifer</td>\n",
       "      <td>Banks</td>\n",
       "      <td>F</td>\n",
       "      <td>561 Perry Cove</td>\n",
       "      <td>...</td>\n",
       "      <td>36.0788</td>\n",
       "      <td>-81.1781</td>\n",
       "      <td>3495</td>\n",
       "      <td>Psychologist, counselling</td>\n",
       "      <td>1988-03-09</td>\n",
       "      <td>0b242abb623afc578575680df30655b9</td>\n",
       "      <td>1325376018</td>\n",
       "      <td>36.011293</td>\n",
       "      <td>-82.048315</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2019-01-01 00:00:44</td>\n",
       "      <td>630423337322</td>\n",
       "      <td>fraud_Heller, Gutmann and Zieme</td>\n",
       "      <td>grocery_pos</td>\n",
       "      <td>107.23</td>\n",
       "      <td>Stephanie</td>\n",
       "      <td>Gill</td>\n",
       "      <td>F</td>\n",
       "      <td>43039 Riley Greens Suite 393</td>\n",
       "      <td>...</td>\n",
       "      <td>48.8878</td>\n",
       "      <td>-118.2105</td>\n",
       "      <td>149</td>\n",
       "      <td>Special educational needs teacher</td>\n",
       "      <td>1978-06-21</td>\n",
       "      <td>1f76529f8574734946361c461b024d99</td>\n",
       "      <td>1325376044</td>\n",
       "      <td>49.159047</td>\n",
       "      <td>-118.186462</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>2019-01-01 00:00:51</td>\n",
       "      <td>38859492057661</td>\n",
       "      <td>fraud_Lind-Buckridge</td>\n",
       "      <td>entertainment</td>\n",
       "      <td>220.11</td>\n",
       "      <td>Edward</td>\n",
       "      <td>Sanchez</td>\n",
       "      <td>M</td>\n",
       "      <td>594 White Dale Suite 530</td>\n",
       "      <td>...</td>\n",
       "      <td>42.1808</td>\n",
       "      <td>-112.2620</td>\n",
       "      <td>4154</td>\n",
       "      <td>Nature conservation officer</td>\n",
       "      <td>1962-01-19</td>\n",
       "      <td>a1a22d70485983eac12b5b88dad1cf95</td>\n",
       "      <td>1325376051</td>\n",
       "      <td>43.150704</td>\n",
       "      <td>-112.154481</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>2019-01-01 00:01:16</td>\n",
       "      <td>3534093764340240</td>\n",
       "      <td>fraud_Kutch, Hermiston and Farrell</td>\n",
       "      <td>gas_transport</td>\n",
       "      <td>45.00</td>\n",
       "      <td>Jeremy</td>\n",
       "      <td>White</td>\n",
       "      <td>M</td>\n",
       "      <td>9443 Cynthia Court Apt. 038</td>\n",
       "      <td>...</td>\n",
       "      <td>46.2306</td>\n",
       "      <td>-112.1138</td>\n",
       "      <td>1939</td>\n",
       "      <td>Patent attorney</td>\n",
       "      <td>1967-01-12</td>\n",
       "      <td>6b849c168bdad6f867558c3793159a81</td>\n",
       "      <td>1325376076</td>\n",
       "      <td>47.034331</td>\n",
       "      <td>-112.561071</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>2019-01-01 00:03:06</td>\n",
       "      <td>375534208663984</td>\n",
       "      <td>fraud_Keeling-Crist</td>\n",
       "      <td>misc_pos</td>\n",
       "      <td>41.96</td>\n",
       "      <td>Tyler</td>\n",
       "      <td>Garcia</td>\n",
       "      <td>M</td>\n",
       "      <td>408 Bradley Rest</td>\n",
       "      <td>...</td>\n",
       "      <td>38.4207</td>\n",
       "      <td>-79.4629</td>\n",
       "      <td>99</td>\n",
       "      <td>Dance movement psychotherapist</td>\n",
       "      <td>1986-03-28</td>\n",
       "      <td>a41d7549acf90789359a9aa5346dcb46</td>\n",
       "      <td>1325376186</td>\n",
       "      <td>38.674999</td>\n",
       "      <td>-78.632459</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296670</th>\n",
       "      <td>1296670</td>\n",
       "      <td>2020-06-21 12:12:08</td>\n",
       "      <td>30263540414123</td>\n",
       "      <td>fraud_Reichel Inc</td>\n",
       "      <td>entertainment</td>\n",
       "      <td>15.56</td>\n",
       "      <td>Erik</td>\n",
       "      <td>Patterson</td>\n",
       "      <td>M</td>\n",
       "      <td>162 Jessica Row Apt. 072</td>\n",
       "      <td>...</td>\n",
       "      <td>37.7175</td>\n",
       "      <td>-112.4777</td>\n",
       "      <td>258</td>\n",
       "      <td>Geoscientist</td>\n",
       "      <td>1961-11-24</td>\n",
       "      <td>440b587732da4dc1a6395aba5fb41669</td>\n",
       "      <td>1371816728</td>\n",
       "      <td>36.841266</td>\n",
       "      <td>-111.690765</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296671</th>\n",
       "      <td>1296671</td>\n",
       "      <td>2020-06-21 12:12:19</td>\n",
       "      <td>6011149206456997</td>\n",
       "      <td>fraud_Abernathy and Sons</td>\n",
       "      <td>food_dining</td>\n",
       "      <td>51.70</td>\n",
       "      <td>Jeffrey</td>\n",
       "      <td>White</td>\n",
       "      <td>M</td>\n",
       "      <td>8617 Holmes Terrace Suite 651</td>\n",
       "      <td>...</td>\n",
       "      <td>39.2667</td>\n",
       "      <td>-77.5101</td>\n",
       "      <td>100</td>\n",
       "      <td>Production assistant, television</td>\n",
       "      <td>1979-12-11</td>\n",
       "      <td>278000d2e0d2277d1de2f890067dcc0a</td>\n",
       "      <td>1371816739</td>\n",
       "      <td>38.906881</td>\n",
       "      <td>-78.246528</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296672</th>\n",
       "      <td>1296672</td>\n",
       "      <td>2020-06-21 12:12:32</td>\n",
       "      <td>3514865930894695</td>\n",
       "      <td>fraud_Stiedemann Ltd</td>\n",
       "      <td>food_dining</td>\n",
       "      <td>105.93</td>\n",
       "      <td>Christopher</td>\n",
       "      <td>Castaneda</td>\n",
       "      <td>M</td>\n",
       "      <td>1632 Cohen Drive Suite 639</td>\n",
       "      <td>...</td>\n",
       "      <td>32.9396</td>\n",
       "      <td>-105.8189</td>\n",
       "      <td>899</td>\n",
       "      <td>Naval architect</td>\n",
       "      <td>1967-08-30</td>\n",
       "      <td>483f52fe67fabef353d552c1e662974c</td>\n",
       "      <td>1371816752</td>\n",
       "      <td>33.619513</td>\n",
       "      <td>-105.130529</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296673</th>\n",
       "      <td>1296673</td>\n",
       "      <td>2020-06-21 12:13:36</td>\n",
       "      <td>2720012583106919</td>\n",
       "      <td>fraud_Reinger, Weissnat and Strosin</td>\n",
       "      <td>food_dining</td>\n",
       "      <td>74.90</td>\n",
       "      <td>Joseph</td>\n",
       "      <td>Murray</td>\n",
       "      <td>M</td>\n",
       "      <td>42933 Ryan Underpass</td>\n",
       "      <td>...</td>\n",
       "      <td>43.3526</td>\n",
       "      <td>-102.5411</td>\n",
       "      <td>1126</td>\n",
       "      <td>Volunteer coordinator</td>\n",
       "      <td>1980-08-18</td>\n",
       "      <td>d667cdcbadaaed3da3f4020e83591c83</td>\n",
       "      <td>1371816816</td>\n",
       "      <td>42.788940</td>\n",
       "      <td>-103.241160</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296674</th>\n",
       "      <td>1296674</td>\n",
       "      <td>2020-06-21 12:13:37</td>\n",
       "      <td>4292902571056973207</td>\n",
       "      <td>fraud_Langosh, Wintheiser and Hyatt</td>\n",
       "      <td>food_dining</td>\n",
       "      <td>4.30</td>\n",
       "      <td>Jeffrey</td>\n",
       "      <td>Smith</td>\n",
       "      <td>M</td>\n",
       "      <td>135 Joseph Mountains</td>\n",
       "      <td>...</td>\n",
       "      <td>45.8433</td>\n",
       "      <td>-113.8748</td>\n",
       "      <td>218</td>\n",
       "      <td>Therapist, horticultural</td>\n",
       "      <td>1995-08-16</td>\n",
       "      <td>8f7c8e4ab7f25875d753b422917c98c9</td>\n",
       "      <td>1371816817</td>\n",
       "      <td>46.565983</td>\n",
       "      <td>-114.186110</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1296675 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         Unnamed: 0 trans_date_trans_time               cc_num  \\\n",
       "0                 0   2019-01-01 00:00:18     2703186189652095   \n",
       "1                 1   2019-01-01 00:00:44         630423337322   \n",
       "2                 2   2019-01-01 00:00:51       38859492057661   \n",
       "3                 3   2019-01-01 00:01:16     3534093764340240   \n",
       "4                 4   2019-01-01 00:03:06      375534208663984   \n",
       "...             ...                   ...                  ...   \n",
       "1296670     1296670   2020-06-21 12:12:08       30263540414123   \n",
       "1296671     1296671   2020-06-21 12:12:19     6011149206456997   \n",
       "1296672     1296672   2020-06-21 12:12:32     3514865930894695   \n",
       "1296673     1296673   2020-06-21 12:13:36     2720012583106919   \n",
       "1296674     1296674   2020-06-21 12:13:37  4292902571056973207   \n",
       "\n",
       "                                    merchant       category     amt  \\\n",
       "0                 fraud_Rippin, Kub and Mann       misc_net    4.97   \n",
       "1            fraud_Heller, Gutmann and Zieme    grocery_pos  107.23   \n",
       "2                       fraud_Lind-Buckridge  entertainment  220.11   \n",
       "3         fraud_Kutch, Hermiston and Farrell  gas_transport   45.00   \n",
       "4                        fraud_Keeling-Crist       misc_pos   41.96   \n",
       "...                                      ...            ...     ...   \n",
       "1296670                    fraud_Reichel Inc  entertainment   15.56   \n",
       "1296671             fraud_Abernathy and Sons    food_dining   51.70   \n",
       "1296672                 fraud_Stiedemann Ltd    food_dining  105.93   \n",
       "1296673  fraud_Reinger, Weissnat and Strosin    food_dining   74.90   \n",
       "1296674  fraud_Langosh, Wintheiser and Hyatt    food_dining    4.30   \n",
       "\n",
       "               first       last gender                         street  ...  \\\n",
       "0           Jennifer      Banks      F                 561 Perry Cove  ...   \n",
       "1          Stephanie       Gill      F   43039 Riley Greens Suite 393  ...   \n",
       "2             Edward    Sanchez      M       594 White Dale Suite 530  ...   \n",
       "3             Jeremy      White      M    9443 Cynthia Court Apt. 038  ...   \n",
       "4              Tyler     Garcia      M               408 Bradley Rest  ...   \n",
       "...              ...        ...    ...                            ...  ...   \n",
       "1296670         Erik  Patterson      M       162 Jessica Row Apt. 072  ...   \n",
       "1296671      Jeffrey      White      M  8617 Holmes Terrace Suite 651  ...   \n",
       "1296672  Christopher  Castaneda      M     1632 Cohen Drive Suite 639  ...   \n",
       "1296673       Joseph     Murray      M           42933 Ryan Underpass  ...   \n",
       "1296674      Jeffrey      Smith      M           135 Joseph Mountains  ...   \n",
       "\n",
       "             lat      long  city_pop                                job  \\\n",
       "0        36.0788  -81.1781      3495          Psychologist, counselling   \n",
       "1        48.8878 -118.2105       149  Special educational needs teacher   \n",
       "2        42.1808 -112.2620      4154        Nature conservation officer   \n",
       "3        46.2306 -112.1138      1939                    Patent attorney   \n",
       "4        38.4207  -79.4629        99     Dance movement psychotherapist   \n",
       "...          ...       ...       ...                                ...   \n",
       "1296670  37.7175 -112.4777       258                       Geoscientist   \n",
       "1296671  39.2667  -77.5101       100   Production assistant, television   \n",
       "1296672  32.9396 -105.8189       899                    Naval architect   \n",
       "1296673  43.3526 -102.5411      1126              Volunteer coordinator   \n",
       "1296674  45.8433 -113.8748       218           Therapist, horticultural   \n",
       "\n",
       "                dob                         trans_num   unix_time  merch_lat  \\\n",
       "0        1988-03-09  0b242abb623afc578575680df30655b9  1325376018  36.011293   \n",
       "1        1978-06-21  1f76529f8574734946361c461b024d99  1325376044  49.159047   \n",
       "2        1962-01-19  a1a22d70485983eac12b5b88dad1cf95  1325376051  43.150704   \n",
       "3        1967-01-12  6b849c168bdad6f867558c3793159a81  1325376076  47.034331   \n",
       "4        1986-03-28  a41d7549acf90789359a9aa5346dcb46  1325376186  38.674999   \n",
       "...             ...                               ...         ...        ...   \n",
       "1296670  1961-11-24  440b587732da4dc1a6395aba5fb41669  1371816728  36.841266   \n",
       "1296671  1979-12-11  278000d2e0d2277d1de2f890067dcc0a  1371816739  38.906881   \n",
       "1296672  1967-08-30  483f52fe67fabef353d552c1e662974c  1371816752  33.619513   \n",
       "1296673  1980-08-18  d667cdcbadaaed3da3f4020e83591c83  1371816816  42.788940   \n",
       "1296674  1995-08-16  8f7c8e4ab7f25875d753b422917c98c9  1371816817  46.565983   \n",
       "\n",
       "         merch_long  is_fraud  \n",
       "0        -82.048315         0  \n",
       "1       -118.186462         0  \n",
       "2       -112.154481         0  \n",
       "3       -112.561071         0  \n",
       "4        -78.632459         0  \n",
       "...             ...       ...  \n",
       "1296670 -111.690765         0  \n",
       "1296671  -78.246528         0  \n",
       "1296672 -105.130529         0  \n",
       "1296673 -103.241160         0  \n",
       "1296674 -114.186110         0  \n",
       "\n",
       "[1296675 rows x 23 columns]"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 1296675 entries, 0 to 1296674\n",
      "Data columns (total 23 columns):\n",
      " #   Column                 Non-Null Count    Dtype  \n",
      "---  ------                 --------------    -----  \n",
      " 0   Unnamed: 0             1296675 non-null  int64  \n",
      " 1   trans_date_trans_time  1296675 non-null  object \n",
      " 2   cc_num                 1296675 non-null  int64  \n",
      " 3   merchant               1296675 non-null  object \n",
      " 4   category               1296675 non-null  object \n",
      " 5   amt                    1296675 non-null  float64\n",
      " 6   first                  1296675 non-null  object \n",
      " 7   last                   1296675 non-null  object \n",
      " 8   gender                 1296675 non-null  object \n",
      " 9   street                 1296675 non-null  object \n",
      " 10  city                   1296675 non-null  object \n",
      " 11  state                  1296675 non-null  object \n",
      " 12  zip                    1296675 non-null  int64  \n",
      " 13  lat                    1296675 non-null  float64\n",
      " 14  long                   1296675 non-null  float64\n",
      " 15  city_pop               1296675 non-null  int64  \n",
      " 16  job                    1296675 non-null  object \n",
      " 17  dob                    1296675 non-null  object \n",
      " 18  trans_num              1296675 non-null  object \n",
      " 19  unix_time              1296675 non-null  int64  \n",
      " 20  merch_lat              1296675 non-null  float64\n",
      " 21  merch_long             1296675 non-null  float64\n",
      " 22  is_fraud               1296675 non-null  int64  \n",
      "dtypes: float64(5), int64(6), object(12)\n",
      "memory usage: 227.5+ MB\n"
     ]
    }
   ],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Unnamed: 0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>cc_num</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>amt</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>zip</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>lat</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>long</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>city_pop</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>unix_time</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>merch_lat</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>merch_long</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>is_fraud</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             0\n",
       "0   Unnamed: 0\n",
       "1       cc_num\n",
       "2          amt\n",
       "3          zip\n",
       "4          lat\n",
       "5         long\n",
       "6     city_pop\n",
       "7    unix_time\n",
       "8    merch_lat\n",
       "9   merch_long\n",
       "10    is_fraud"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "num_columns = data.select_dtypes(include='number').columns\n",
    "pd.DataFrame(num_columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "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>cc_num</th>\n",
       "      <th>amt</th>\n",
       "      <th>zip</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "      <th>city_pop</th>\n",
       "      <th>unix_time</th>\n",
       "      <th>merch_lat</th>\n",
       "      <th>merch_long</th>\n",
       "      <th>is_fraud</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2703186189652095</td>\n",
       "      <td>4.97</td>\n",
       "      <td>28654</td>\n",
       "      <td>36.0788</td>\n",
       "      <td>-81.1781</td>\n",
       "      <td>3495</td>\n",
       "      <td>1325376018</td>\n",
       "      <td>36.011293</td>\n",
       "      <td>-82.048315</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>630423337322</td>\n",
       "      <td>107.23</td>\n",
       "      <td>99160</td>\n",
       "      <td>48.8878</td>\n",
       "      <td>-118.2105</td>\n",
       "      <td>149</td>\n",
       "      <td>1325376044</td>\n",
       "      <td>49.159047</td>\n",
       "      <td>-118.186462</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38859492057661</td>\n",
       "      <td>220.11</td>\n",
       "      <td>83252</td>\n",
       "      <td>42.1808</td>\n",
       "      <td>-112.2620</td>\n",
       "      <td>4154</td>\n",
       "      <td>1325376051</td>\n",
       "      <td>43.150704</td>\n",
       "      <td>-112.154481</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3534093764340240</td>\n",
       "      <td>45.00</td>\n",
       "      <td>59632</td>\n",
       "      <td>46.2306</td>\n",
       "      <td>-112.1138</td>\n",
       "      <td>1939</td>\n",
       "      <td>1325376076</td>\n",
       "      <td>47.034331</td>\n",
       "      <td>-112.561071</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>375534208663984</td>\n",
       "      <td>41.96</td>\n",
       "      <td>24433</td>\n",
       "      <td>38.4207</td>\n",
       "      <td>-79.4629</td>\n",
       "      <td>99</td>\n",
       "      <td>1325376186</td>\n",
       "      <td>38.674999</td>\n",
       "      <td>-78.632459</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296670</th>\n",
       "      <td>30263540414123</td>\n",
       "      <td>15.56</td>\n",
       "      <td>84735</td>\n",
       "      <td>37.7175</td>\n",
       "      <td>-112.4777</td>\n",
       "      <td>258</td>\n",
       "      <td>1371816728</td>\n",
       "      <td>36.841266</td>\n",
       "      <td>-111.690765</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296671</th>\n",
       "      <td>6011149206456997</td>\n",
       "      <td>51.70</td>\n",
       "      <td>21790</td>\n",
       "      <td>39.2667</td>\n",
       "      <td>-77.5101</td>\n",
       "      <td>100</td>\n",
       "      <td>1371816739</td>\n",
       "      <td>38.906881</td>\n",
       "      <td>-78.246528</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296672</th>\n",
       "      <td>3514865930894695</td>\n",
       "      <td>105.93</td>\n",
       "      <td>88325</td>\n",
       "      <td>32.9396</td>\n",
       "      <td>-105.8189</td>\n",
       "      <td>899</td>\n",
       "      <td>1371816752</td>\n",
       "      <td>33.619513</td>\n",
       "      <td>-105.130529</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296673</th>\n",
       "      <td>2720012583106919</td>\n",
       "      <td>74.90</td>\n",
       "      <td>57756</td>\n",
       "      <td>43.3526</td>\n",
       "      <td>-102.5411</td>\n",
       "      <td>1126</td>\n",
       "      <td>1371816816</td>\n",
       "      <td>42.788940</td>\n",
       "      <td>-103.241160</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296674</th>\n",
       "      <td>4292902571056973207</td>\n",
       "      <td>4.30</td>\n",
       "      <td>59871</td>\n",
       "      <td>45.8433</td>\n",
       "      <td>-113.8748</td>\n",
       "      <td>218</td>\n",
       "      <td>1371816817</td>\n",
       "      <td>46.565983</td>\n",
       "      <td>-114.186110</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1296675 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                      cc_num     amt    zip      lat      long  city_pop  \\\n",
       "0           2703186189652095    4.97  28654  36.0788  -81.1781      3495   \n",
       "1               630423337322  107.23  99160  48.8878 -118.2105       149   \n",
       "2             38859492057661  220.11  83252  42.1808 -112.2620      4154   \n",
       "3           3534093764340240   45.00  59632  46.2306 -112.1138      1939   \n",
       "4            375534208663984   41.96  24433  38.4207  -79.4629        99   \n",
       "...                      ...     ...    ...      ...       ...       ...   \n",
       "1296670       30263540414123   15.56  84735  37.7175 -112.4777       258   \n",
       "1296671     6011149206456997   51.70  21790  39.2667  -77.5101       100   \n",
       "1296672     3514865930894695  105.93  88325  32.9396 -105.8189       899   \n",
       "1296673     2720012583106919   74.90  57756  43.3526 -102.5411      1126   \n",
       "1296674  4292902571056973207    4.30  59871  45.8433 -113.8748       218   \n",
       "\n",
       "          unix_time  merch_lat  merch_long  is_fraud  \n",
       "0        1325376018  36.011293  -82.048315         0  \n",
       "1        1325376044  49.159047 -118.186462         0  \n",
       "2        1325376051  43.150704 -112.154481         0  \n",
       "3        1325376076  47.034331 -112.561071         0  \n",
       "4        1325376186  38.674999  -78.632459         0  \n",
       "...             ...        ...         ...       ...  \n",
       "1296670  1371816728  36.841266 -111.690765         0  \n",
       "1296671  1371816739  38.906881  -78.246528         0  \n",
       "1296672  1371816752  33.619513 -105.130529         0  \n",
       "1296673  1371816816  42.788940 -103.241160         0  \n",
       "1296674  1371816817  46.565983 -114.186110         0  \n",
       "\n",
       "[1296675 rows x 10 columns]"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = data[num_columns].copy()\n",
    "data.drop(\"Unnamed: 0\",axis=1,inplace=True)\n",
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "1kBwtX8L-4nl"
   },
   "source": [
    "# Credit Card Fraud Detection - Artificial Neural Networks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "executionInfo": {
     "elapsed": 293,
     "status": "ok",
     "timestamp": 1734381085848,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "Wb1hyDPq4i5M"
   },
   "outputs": [],
   "source": [
    "features_train = data.iloc[:, :-1]\n",
    "targets_train = data.iloc[:,-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 290,
     "status": "ok",
     "timestamp": 1734381086990,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "mqDbIQw-6X4O",
    "outputId": "8e0d859b-cfa1-4eac-d233-b6f7b4b9dd38"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1296675, 9)\n",
      "(1296675,)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>cc_num</th>\n",
       "      <th>amt</th>\n",
       "      <th>zip</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "      <th>city_pop</th>\n",
       "      <th>unix_time</th>\n",
       "      <th>merch_lat</th>\n",
       "      <th>merch_long</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2703186189652095</td>\n",
       "      <td>4.97</td>\n",
       "      <td>28654</td>\n",
       "      <td>36.0788</td>\n",
       "      <td>-81.1781</td>\n",
       "      <td>3495</td>\n",
       "      <td>1325376018</td>\n",
       "      <td>36.011293</td>\n",
       "      <td>-82.048315</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>630423337322</td>\n",
       "      <td>107.23</td>\n",
       "      <td>99160</td>\n",
       "      <td>48.8878</td>\n",
       "      <td>-118.2105</td>\n",
       "      <td>149</td>\n",
       "      <td>1325376044</td>\n",
       "      <td>49.159047</td>\n",
       "      <td>-118.186462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38859492057661</td>\n",
       "      <td>220.11</td>\n",
       "      <td>83252</td>\n",
       "      <td>42.1808</td>\n",
       "      <td>-112.2620</td>\n",
       "      <td>4154</td>\n",
       "      <td>1325376051</td>\n",
       "      <td>43.150704</td>\n",
       "      <td>-112.154481</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3534093764340240</td>\n",
       "      <td>45.00</td>\n",
       "      <td>59632</td>\n",
       "      <td>46.2306</td>\n",
       "      <td>-112.1138</td>\n",
       "      <td>1939</td>\n",
       "      <td>1325376076</td>\n",
       "      <td>47.034331</td>\n",
       "      <td>-112.561071</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>375534208663984</td>\n",
       "      <td>41.96</td>\n",
       "      <td>24433</td>\n",
       "      <td>38.4207</td>\n",
       "      <td>-79.4629</td>\n",
       "      <td>99</td>\n",
       "      <td>1325376186</td>\n",
       "      <td>38.674999</td>\n",
       "      <td>-78.632459</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296670</th>\n",
       "      <td>30263540414123</td>\n",
       "      <td>15.56</td>\n",
       "      <td>84735</td>\n",
       "      <td>37.7175</td>\n",
       "      <td>-112.4777</td>\n",
       "      <td>258</td>\n",
       "      <td>1371816728</td>\n",
       "      <td>36.841266</td>\n",
       "      <td>-111.690765</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296671</th>\n",
       "      <td>6011149206456997</td>\n",
       "      <td>51.70</td>\n",
       "      <td>21790</td>\n",
       "      <td>39.2667</td>\n",
       "      <td>-77.5101</td>\n",
       "      <td>100</td>\n",
       "      <td>1371816739</td>\n",
       "      <td>38.906881</td>\n",
       "      <td>-78.246528</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296672</th>\n",
       "      <td>3514865930894695</td>\n",
       "      <td>105.93</td>\n",
       "      <td>88325</td>\n",
       "      <td>32.9396</td>\n",
       "      <td>-105.8189</td>\n",
       "      <td>899</td>\n",
       "      <td>1371816752</td>\n",
       "      <td>33.619513</td>\n",
       "      <td>-105.130529</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296673</th>\n",
       "      <td>2720012583106919</td>\n",
       "      <td>74.90</td>\n",
       "      <td>57756</td>\n",
       "      <td>43.3526</td>\n",
       "      <td>-102.5411</td>\n",
       "      <td>1126</td>\n",
       "      <td>1371816816</td>\n",
       "      <td>42.788940</td>\n",
       "      <td>-103.241160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296674</th>\n",
       "      <td>4292902571056973207</td>\n",
       "      <td>4.30</td>\n",
       "      <td>59871</td>\n",
       "      <td>45.8433</td>\n",
       "      <td>-113.8748</td>\n",
       "      <td>218</td>\n",
       "      <td>1371816817</td>\n",
       "      <td>46.565983</td>\n",
       "      <td>-114.186110</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1296675 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                      cc_num     amt    zip      lat      long  city_pop  \\\n",
       "0           2703186189652095    4.97  28654  36.0788  -81.1781      3495   \n",
       "1               630423337322  107.23  99160  48.8878 -118.2105       149   \n",
       "2             38859492057661  220.11  83252  42.1808 -112.2620      4154   \n",
       "3           3534093764340240   45.00  59632  46.2306 -112.1138      1939   \n",
       "4            375534208663984   41.96  24433  38.4207  -79.4629        99   \n",
       "...                      ...     ...    ...      ...       ...       ...   \n",
       "1296670       30263540414123   15.56  84735  37.7175 -112.4777       258   \n",
       "1296671     6011149206456997   51.70  21790  39.2667  -77.5101       100   \n",
       "1296672     3514865930894695  105.93  88325  32.9396 -105.8189       899   \n",
       "1296673     2720012583106919   74.90  57756  43.3526 -102.5411      1126   \n",
       "1296674  4292902571056973207    4.30  59871  45.8433 -113.8748       218   \n",
       "\n",
       "          unix_time  merch_lat  merch_long  \n",
       "0        1325376018  36.011293  -82.048315  \n",
       "1        1325376044  49.159047 -118.186462  \n",
       "2        1325376051  43.150704 -112.154481  \n",
       "3        1325376076  47.034331 -112.561071  \n",
       "4        1325376186  38.674999  -78.632459  \n",
       "...             ...        ...         ...  \n",
       "1296670  1371816728  36.841266 -111.690765  \n",
       "1296671  1371816739  38.906881  -78.246528  \n",
       "1296672  1371816752  33.619513 -105.130529  \n",
       "1296673  1371816816  42.788940 -103.241160  \n",
       "1296674  1371816817  46.565983 -114.186110  \n",
       "\n",
       "[1296675 rows x 9 columns]"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(features_train.shape)\n",
    "print(targets_train.shape)\n",
    "features_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "is_fraud\n",
       "0    1289169\n",
       "1       7506\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "targets_train.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "executionInfo": {
     "elapsed": 274,
     "status": "ok",
     "timestamp": 1734381095444,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "V5vrl19T6wr7"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "scaler = StandardScaler()\n",
    "\n",
    "features_train_sclr = pd.DataFrame(\n",
    "    scaler.fit_transform(features_train),\n",
    "    columns=features_train.columns\n",
    ")\n",
    "\n",
    "features_train_sclr.shape[1]"
   ]
  },
  {
   "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>cc_num</th>\n",
       "      <th>amt</th>\n",
       "      <th>zip</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "      <th>city_pop</th>\n",
       "      <th>unix_time</th>\n",
       "      <th>merch_lat</th>\n",
       "      <th>merch_long</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.316692</td>\n",
       "      <td>-0.407826</td>\n",
       "      <td>-0.749136</td>\n",
       "      <td>-0.484420</td>\n",
       "      <td>0.657620</td>\n",
       "      <td>-0.282589</td>\n",
       "      <td>-1.858664</td>\n",
       "      <td>-0.494354</td>\n",
       "      <td>0.593864</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.318757</td>\n",
       "      <td>0.230039</td>\n",
       "      <td>1.872567</td>\n",
       "      <td>2.039120</td>\n",
       "      <td>-2.033870</td>\n",
       "      <td>-0.293670</td>\n",
       "      <td>-1.858662</td>\n",
       "      <td>2.078699</td>\n",
       "      <td>-2.030341</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.318728</td>\n",
       "      <td>0.934149</td>\n",
       "      <td>1.281042</td>\n",
       "      <td>0.717754</td>\n",
       "      <td>-1.601537</td>\n",
       "      <td>-0.280406</td>\n",
       "      <td>-1.858662</td>\n",
       "      <td>0.902849</td>\n",
       "      <td>-1.592323</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.316058</td>\n",
       "      <td>-0.158132</td>\n",
       "      <td>0.402753</td>\n",
       "      <td>1.515617</td>\n",
       "      <td>-1.590766</td>\n",
       "      <td>-0.287742</td>\n",
       "      <td>-1.858660</td>\n",
       "      <td>1.662886</td>\n",
       "      <td>-1.621848</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-0.318471</td>\n",
       "      <td>-0.177094</td>\n",
       "      <td>-0.906090</td>\n",
       "      <td>-0.023035</td>\n",
       "      <td>0.782279</td>\n",
       "      <td>-0.293835</td>\n",
       "      <td>-1.858651</td>\n",
       "      <td>0.026941</td>\n",
       "      <td>0.841909</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296670</th>\n",
       "      <td>-0.318735</td>\n",
       "      <td>-0.341769</td>\n",
       "      <td>1.336186</td>\n",
       "      <td>-0.161575</td>\n",
       "      <td>-1.617214</td>\n",
       "      <td>-0.293309</td>\n",
       "      <td>1.757855</td>\n",
       "      <td>-0.331926</td>\n",
       "      <td>-1.558650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296671</th>\n",
       "      <td>-0.314165</td>\n",
       "      <td>-0.116339</td>\n",
       "      <td>-1.004368</td>\n",
       "      <td>0.143638</td>\n",
       "      <td>0.924207</td>\n",
       "      <td>-0.293832</td>\n",
       "      <td>1.757855</td>\n",
       "      <td>0.072321</td>\n",
       "      <td>0.869934</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296672</th>\n",
       "      <td>-0.316072</td>\n",
       "      <td>0.221930</td>\n",
       "      <td>1.469677</td>\n",
       "      <td>-1.102883</td>\n",
       "      <td>-1.133257</td>\n",
       "      <td>-0.291186</td>\n",
       "      <td>1.757856</td>\n",
       "      <td>-0.962433</td>\n",
       "      <td>-1.082272</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296673</th>\n",
       "      <td>-0.316680</td>\n",
       "      <td>0.028375</td>\n",
       "      <td>0.332996</td>\n",
       "      <td>0.948613</td>\n",
       "      <td>-0.895029</td>\n",
       "      <td>-0.290434</td>\n",
       "      <td>1.757861</td>\n",
       "      <td>0.832051</td>\n",
       "      <td>-0.945074</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1296674</th>\n",
       "      <td>2.961257</td>\n",
       "      <td>-0.412005</td>\n",
       "      <td>0.411640</td>\n",
       "      <td>1.439314</td>\n",
       "      <td>-1.718754</td>\n",
       "      <td>-0.293441</td>\n",
       "      <td>1.757862</td>\n",
       "      <td>1.571229</td>\n",
       "      <td>-1.739852</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1296675 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           cc_num       amt       zip       lat      long  city_pop  \\\n",
       "0       -0.316692 -0.407826 -0.749136 -0.484420  0.657620 -0.282589   \n",
       "1       -0.318757  0.230039  1.872567  2.039120 -2.033870 -0.293670   \n",
       "2       -0.318728  0.934149  1.281042  0.717754 -1.601537 -0.280406   \n",
       "3       -0.316058 -0.158132  0.402753  1.515617 -1.590766 -0.287742   \n",
       "4       -0.318471 -0.177094 -0.906090 -0.023035  0.782279 -0.293835   \n",
       "...           ...       ...       ...       ...       ...       ...   \n",
       "1296670 -0.318735 -0.341769  1.336186 -0.161575 -1.617214 -0.293309   \n",
       "1296671 -0.314165 -0.116339 -1.004368  0.143638  0.924207 -0.293832   \n",
       "1296672 -0.316072  0.221930  1.469677 -1.102883 -1.133257 -0.291186   \n",
       "1296673 -0.316680  0.028375  0.332996  0.948613 -0.895029 -0.290434   \n",
       "1296674  2.961257 -0.412005  0.411640  1.439314 -1.718754 -0.293441   \n",
       "\n",
       "         unix_time  merch_lat  merch_long  \n",
       "0        -1.858664  -0.494354    0.593864  \n",
       "1        -1.858662   2.078699   -2.030341  \n",
       "2        -1.858662   0.902849   -1.592323  \n",
       "3        -1.858660   1.662886   -1.621848  \n",
       "4        -1.858651   0.026941    0.841909  \n",
       "...            ...        ...         ...  \n",
       "1296670   1.757855  -0.331926   -1.558650  \n",
       "1296671   1.757855   0.072321    0.869934  \n",
       "1296672   1.757856  -0.962433   -1.082272  \n",
       "1296673   1.757861   0.832051   -0.945074  \n",
       "1296674   1.757862   1.571229   -1.739852  \n",
       "\n",
       "[1296675 rows x 9 columns]"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features_train_sclr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {
    "executionInfo": {
     "elapsed": 259,
     "status": "ok",
     "timestamp": 1734381518217,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "3B5vHzsA6-F9"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential_11\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"sequential_11\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> Layer (type)                    </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">       Param # </span>┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ dense_44 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)             │           <span style=\"color: #00af00; text-decoration-color: #00af00\">640</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_33 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)            │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)             │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_45 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>)             │         <span style=\"color: #00af00; text-decoration-color: #00af00\">2,080</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_34 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)            │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>)             │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_46 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>)             │           <span style=\"color: #00af00; text-decoration-color: #00af00\">528</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_35 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)            │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>)             │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_47 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)              │            <span style=\"color: #00af00; text-decoration-color: #00af00\">17</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                   \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m      Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ dense_44 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m64\u001b[0m)             │           \u001b[38;5;34m640\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_33 (\u001b[38;5;33mDropout\u001b[0m)            │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m64\u001b[0m)             │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_45 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m)             │         \u001b[38;5;34m2,080\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_34 (\u001b[38;5;33mDropout\u001b[0m)            │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m)             │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_46 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m16\u001b[0m)             │           \u001b[38;5;34m528\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_35 (\u001b[38;5;33mDropout\u001b[0m)            │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m16\u001b[0m)             │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_47 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)              │            \u001b[38;5;34m17\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">3,265</span> (12.75 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m3,265\u001b[0m (12.75 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">3,265</span> (12.75 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m3,265\u001b[0m (12.75 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from keras import models, layers, regularizers, optimizers\n",
    "\n",
    "def create_model(input_shape):\n",
    "    model = models.Sequential([\n",
    "        layers.Input(shape=input_shape),  # Use Input layer for the first layer\n",
    "        layers.Dense(64, activation='relu', kernel_regularizer=regularizers.l2(0.001)),\n",
    "        layers.Dropout(0.5),\n",
    "        layers.Dense(32, activation='relu', kernel_regularizer=regularizers.l2(0.001)),\n",
    "        layers.Dropout(0.5),\n",
    "        layers.Dense(16, activation='relu'),\n",
    "        layers.Dropout(0.5),\n",
    "        layers.Dense(1, activation='sigmoid')\n",
    "    ])\n",
    "    model.compile(optimizer=optimizers.Adam(learning_rate=0.001), \n",
    "                  loss='binary_crossentropy', \n",
    "                  metrics=['accuracy'])\n",
    "    return model\n",
    "\n",
    "model = create_model((features_train.shape[1],))  # Novo modelo a cada iteração\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## SMOTE - Oversampling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\danrl\\miniconda3\\envs\\neural_env\\lib\\site-packages\\sklearn\\base.py:474: FutureWarning: `BaseEstimator._validate_data` is deprecated in 1.6 and will be removed in 1.7. Use `sklearn.utils.validation.validate_data` instead. This function becomes public and is part of the scikit-learn developer API.\n",
      "  warnings.warn(\n",
      "c:\\Users\\danrl\\miniconda3\\envs\\neural_env\\lib\\site-packages\\sklearn\\utils\\_tags.py:354: FutureWarning: The SMOTE or classes from which it inherits use `_get_tags` and `_more_tags`. Please define the `__sklearn_tags__` method, or inherit from `sklearn.base.BaseEstimator` and/or other appropriate mixins such as `sklearn.base.TransformerMixin`, `sklearn.base.ClassifierMixin`, `sklearn.base.RegressorMixin`, and `sklearn.base.OutlierMixin`. From scikit-learn 1.7, not defining `__sklearn_tags__` will raise an error.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tamanho original de X_train: (1296675, 9)\n",
      "Tamanho após SMOTE: (2578338, 9)\n"
     ]
    }
   ],
   "source": [
    "from imblearn.over_sampling import SMOTE\n",
    "\n",
    "# Aplicando SMOTE para oversampling da classe minoritária\n",
    "smote = SMOTE(sampling_strategy='auto', random_state=42)\n",
    "X_train_res, y_train_res = smote.fit_resample(features_train_sclr, targets_train)\n",
    "\n",
    "print(f\"Tamanho original de X_train: {features_train_sclr.shape}\")\n",
    "print(f\"Tamanho após SMOTE: {X_train_res.shape}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "is_fraud\n",
       "0    1289169\n",
       "1    1289169\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_res.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "features_train_sclr = X_train_res\n",
    "targets_train = y_train_res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "is_fraud\n",
       "0    1289169\n",
       "1    1289169\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "targets_train.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Cross-Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 2ms/step - accuracy: 0.8331 - loss: 0.4320 - val_accuracy: 0.8639 - val_loss: 0.3115 - learning_rate: 0.0010\n",
      "Epoch 2/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m26s\u001b[0m 3ms/step - accuracy: 0.8591 - loss: 0.3508 - val_accuracy: 0.8809 - val_loss: 0.2982 - learning_rate: 0.0010\n",
      "Epoch 3/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m22s\u001b[0m 3ms/step - accuracy: 0.8619 - loss: 0.3419 - val_accuracy: 0.8807 - val_loss: 0.2877 - learning_rate: 0.0010\n",
      "Epoch 4/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m25s\u001b[0m 3ms/step - accuracy: 0.8634 - loss: 0.3354 - val_accuracy: 0.8850 - val_loss: 0.2852 - learning_rate: 0.0010\n",
      "Epoch 5/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m27s\u001b[0m 3ms/step - accuracy: 0.8633 - loss: 0.3341 - val_accuracy: 0.8911 - val_loss: 0.2821 - learning_rate: 0.0010\n",
      "Fold 1 - Loss: 0.28209564089775085 - Accuracy: 0.8910733461380005\n",
      "\u001b[1m16115/16115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m24s\u001b[0m 2ms/step\n",
      "Fold 1 - Precision: 0.8662 - Recall: 0.9251 - F1 Score: 0.8947\n",
      "Epoch 1/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m86s\u001b[0m 10ms/step - accuracy: 0.8361 - loss: 0.4297 - val_accuracy: 0.8645 - val_loss: 0.3153 - learning_rate: 0.0010\n",
      "Epoch 2/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m43s\u001b[0m 5ms/step - accuracy: 0.8605 - loss: 0.3478 - val_accuracy: 0.8790 - val_loss: 0.2944 - learning_rate: 0.0010\n",
      "Epoch 3/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m37s\u001b[0m 5ms/step - accuracy: 0.8622 - loss: 0.3407 - val_accuracy: 0.8874 - val_loss: 0.2908 - learning_rate: 0.0010\n",
      "Epoch 4/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m36s\u001b[0m 4ms/step - accuracy: 0.8642 - loss: 0.3365 - val_accuracy: 0.8819 - val_loss: 0.2817 - learning_rate: 0.0010\n",
      "Epoch 5/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m34s\u001b[0m 4ms/step - accuracy: 0.8638 - loss: 0.3321 - val_accuracy: 0.8880 - val_loss: 0.2863 - learning_rate: 0.0010\n",
      "Fold 2 - Loss: 0.28168174624443054 - Accuracy: 0.8819337487220764\n",
      "\u001b[1m16115/16115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 1ms/step\n",
      "Fold 2 - Precision: 0.9166 - Recall: 0.8404 - F1 Score: 0.8768\n",
      "Epoch 1/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m38s\u001b[0m 4ms/step - accuracy: 0.8347 - loss: 0.4351 - val_accuracy: 0.8644 - val_loss: 0.3208 - learning_rate: 0.0010\n",
      "Epoch 2/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m74s\u001b[0m 9ms/step - accuracy: 0.8608 - loss: 0.3536 - val_accuracy: 0.8648 - val_loss: 0.3160 - learning_rate: 0.0010\n",
      "Epoch 3/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m122s\u001b[0m 15ms/step - accuracy: 0.8614 - loss: 0.3479 - val_accuracy: 0.8648 - val_loss: 0.3126 - learning_rate: 0.0010\n",
      "Epoch 4/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m50s\u001b[0m 6ms/step - accuracy: 0.8612 - loss: 0.3467 - val_accuracy: 0.8643 - val_loss: 0.3130 - learning_rate: 0.0010\n",
      "Epoch 5/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m55s\u001b[0m 7ms/step - accuracy: 0.8607 - loss: 0.3470 - val_accuracy: 0.8647 - val_loss: 0.3116 - learning_rate: 0.0010\n",
      "Fold 3 - Loss: 0.31163519620895386 - Accuracy: 0.8647463321685791\n",
      "\u001b[1m16115/16115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m38s\u001b[0m 2ms/step\n",
      "Fold 3 - Precision: 0.9647 - Recall: 0.7572 - F1 Score: 0.8484\n",
      "Epoch 1/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m71s\u001b[0m 8ms/step - accuracy: 0.8330 - loss: 0.4364 - val_accuracy: 0.8654 - val_loss: 0.3205 - learning_rate: 0.0010\n",
      "Epoch 2/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m37s\u001b[0m 5ms/step - accuracy: 0.8596 - loss: 0.3541 - val_accuracy: 0.8855 - val_loss: 0.2939 - learning_rate: 0.0010\n",
      "Epoch 3/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m40s\u001b[0m 5ms/step - accuracy: 0.8616 - loss: 0.3442 - val_accuracy: 0.8907 - val_loss: 0.2896 - learning_rate: 0.0010\n",
      "Epoch 4/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m57s\u001b[0m 7ms/step - accuracy: 0.8641 - loss: 0.3376 - val_accuracy: 0.8890 - val_loss: 0.2838 - learning_rate: 0.0010\n",
      "Epoch 5/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m40s\u001b[0m 5ms/step - accuracy: 0.8637 - loss: 0.3333 - val_accuracy: 0.8917 - val_loss: 0.2795 - learning_rate: 0.0010\n",
      "Fold 4 - Loss: 0.27945202589035034 - Accuracy: 0.8917092084884644\n",
      "\u001b[1m16115/16115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m37s\u001b[0m 2ms/step\n",
      "Fold 4 - Precision: 0.8870 - Recall: 0.8978 - F1 Score: 0.8924\n",
      "Epoch 1/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m69s\u001b[0m 8ms/step - accuracy: 0.8316 - loss: 0.4364 - val_accuracy: 0.8643 - val_loss: 0.3225 - learning_rate: 0.0010\n",
      "Epoch 2/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m40s\u001b[0m 5ms/step - accuracy: 0.8604 - loss: 0.3518 - val_accuracy: 0.8644 - val_loss: 0.3007 - learning_rate: 0.0010\n",
      "Epoch 3/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m43s\u001b[0m 5ms/step - accuracy: 0.8601 - loss: 0.3428 - val_accuracy: 0.8645 - val_loss: 0.2928 - learning_rate: 0.0010\n",
      "Epoch 4/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m47s\u001b[0m 6ms/step - accuracy: 0.8613 - loss: 0.3396 - val_accuracy: 0.8642 - val_loss: 0.2931 - learning_rate: 0.0010\n",
      "Epoch 5/5\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m46s\u001b[0m 6ms/step - accuracy: 0.8616 - loss: 0.3356 - val_accuracy: 0.8646 - val_loss: 0.2840 - learning_rate: 0.0010\n",
      "Fold 5 - Loss: 0.2840222716331482 - Accuracy: 0.8646044731140137\n",
      "\u001b[1m16115/16115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m19s\u001b[0m 1ms/step\n",
      "Fold 5 - Precision: 0.9662 - Recall: 0.7556 - F1 Score: 0.8480\n",
      "\n",
      "Resultados da Validação Cruzada:\n",
      "Acurácia média: 0.8788\n",
      "Perda média: 0.2878\n",
      "Precisão média: 0.9201\n",
      "Recall médio: 0.8352\n",
      "F1 Score médio: 0.8721\n"
     ]
    }
   ],
   "source": [
    "from keras import models, layers\n",
    "from keras.utils import to_categorical\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "from sklearn.model_selection import KFold\n",
    "from keras.callbacks import EarlyStopping, ReduceLROnPlateau, ModelCheckpoint\n",
    "from sklearn.metrics import precision_score, recall_score, f1_score\n",
    "from sklearn.utils.class_weight import compute_class_weight\n",
    "\n",
    "# Calculando os pesos das classes\n",
    "#class_weights = compute_class_weight('balanced', classes=[0, 1], y=targets_train)      \n",
    "#class_weight_dict = {0: class_weights[0], 1: class_weights[1]}                     \n",
    "# class_weight_dict = {0: 1., 1: 100.}                class_weight=class_weight_dict,\n",
    "\n",
    "# Callbacks:\n",
    "early_stopping = EarlyStopping(monitor='val_loss', patience=5, restore_best_weights=True)\n",
    "lr_scheduler = ReduceLROnPlateau(monitor='val_loss', factor=0.1, patience=3)\n",
    "checkpoint = ModelCheckpoint(\n",
    "    'model_checkpoint.keras',  # Salva no formato .keras\n",
    "    save_best_only=True,\n",
    "    monitor='val_loss',\n",
    "    mode='min'\n",
    ")\n",
    "\n",
    "# Cross-Validation settings\n",
    "kfold = StratifiedKFold(n_splits=5, shuffle=True, random_state=42) # The folds are made by preserving the percentage of samples for each class.\n",
    "fold_no = 1\n",
    "accuracy_per_fold = []\n",
    "loss_per_fold = []\n",
    "precision_per_fold = []\n",
    "recall_per_fold = []\n",
    "f1_score_per_fold = []\n",
    "\n",
    "# Looping to folds\n",
    "for train_index, val_index in kfold.split(features_train_sclr, targets_train):\n",
    "    # Spreanding the data into training and validation\n",
    "    X_train, X_val = features_train_sclr.iloc[train_index], features_train_sclr.iloc[val_index]\n",
    "    y_train, y_val = targets_train.iloc[train_index], targets_train.iloc[val_index]\n",
    "\n",
    "    # Creating the model for each fold\n",
    "    model = create_model((X_train.shape[1],)) # Calling the model created before\n",
    "\n",
    "    # Training the model\n",
    "    history = model.fit(\n",
    "        X_train, y_train,\n",
    "        epochs=5,\n",
    "        batch_size=256,\n",
    "        validation_data=(X_val, y_val),\n",
    "        verbose=1, # Setting the model fitting visualization\n",
    "        callbacks=[early_stopping, lr_scheduler, checkpoint]\n",
    "    )\n",
    "\n",
    "    # Evaluating the model to this fold\n",
    "    scores = model.evaluate(X_val, y_val, verbose=0)\n",
    "    print(f\"Fold {fold_no} - Loss: {scores[0]} - Accuracy: {scores[1]}\")\n",
    "    accuracy_per_fold.append(scores[1])\n",
    "    loss_per_fold.append(scores[0])\n",
    "\n",
    "    # Predicting on validation set\n",
    "    y_val_pred = (model.predict(X_val) > 0.5).astype(\"int32\")  # Binarize predictions\n",
    "\n",
    "    # Calculating Precision, Recall, and F1 Score for this fold\n",
    "    precision = precision_score(y_val, y_val_pred)\n",
    "    recall = recall_score(y_val, y_val_pred)\n",
    "    f1 = f1_score(y_val, y_val_pred)\n",
    "\n",
    "    precision_per_fold.append(precision)\n",
    "    recall_per_fold.append(recall)\n",
    "    f1_score_per_fold.append(f1)\n",
    "\n",
    "    print(f\"Fold {fold_no} - Precision: {precision:.4f} - Recall: {recall:.4f} - F1 Score: {f1:.4f}\")\n",
    "\n",
    "    fold_no += 1\n",
    "\n",
    "# Final Results\n",
    "print(\"\\nResultados da Validação Cruzada:\")\n",
    "print(f\"Acurácia média: {np.mean(accuracy_per_fold):.4f}\")\n",
    "print(f\"Perda média: {np.mean(loss_per_fold):.4f}\")\n",
    "print(f\"Precisão média: {np.mean(precision_per_fold):.4f}\")\n",
    "print(f\"Recall médio: {np.mean(recall_per_fold):.4f}\")\n",
    "print(f\"F1 Score médio: {np.mean(f1_score_per_fold):.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2578338, 9)"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features_train_sclr.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "is_fraud\n",
       "0    1289169\n",
       "1    1289169\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "targets_train.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Final Model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "data_train_sclr, data_validation_sclr, train_targets, validation_targets = train_test_split(features_train_sclr, targets_train, test_size=.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m22s\u001b[0m 3ms/step - accuracy: 0.8346 - loss: 0.4301\n",
      "Epoch 2/10\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m25s\u001b[0m 3ms/step - accuracy: 0.8588 - loss: 0.3507\n",
      "Epoch 3/10\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m31s\u001b[0m 4ms/step - accuracy: 0.8605 - loss: 0.3421\n",
      "Epoch 4/10\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m55s\u001b[0m 5ms/step - accuracy: 0.8610 - loss: 0.3362\n",
      "Epoch 5/10\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m25s\u001b[0m 3ms/step - accuracy: 0.8607 - loss: 0.3347\n",
      "Epoch 6/10\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m24s\u001b[0m 3ms/step - accuracy: 0.8612 - loss: 0.3325\n",
      "Epoch 7/10\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m22s\u001b[0m 3ms/step - accuracy: 0.8619 - loss: 0.3315\n",
      "Epoch 8/10\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m21s\u001b[0m 3ms/step - accuracy: 0.8619 - loss: 0.3329\n",
      "Epoch 9/10\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m33s\u001b[0m 4ms/step - accuracy: 0.8623 - loss: 0.3317\n",
      "Epoch 10/10\n",
      "\u001b[1m8058/8058\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m30s\u001b[0m 4ms/step - accuracy: 0.8628 - loss: 0.3309\n"
     ]
    }
   ],
   "source": [
    "from keras.callbacks import EarlyStopping, ReduceLROnPlateau, ModelCheckpoint\n",
    "\n",
    "# Treinando o modelo final no conjunto completo de dados\n",
    "final_model = create_model((data_train_sclr.shape[1],))\n",
    "\n",
    "# Callbacks:\n",
    "#early_stopping = EarlyStopping(monitor='val_loss', patience=5, restore_best_weights=True)\n",
    "#lr_scheduler = ReduceLROnPlateau(monitor='val_loss', factor=0.1, patience=3)\n",
    "# checkpoint = ModelCheckpoint('model_checkpoint.keras', save_best_only=True, monitor='val_loss', mode='min')\n",
    "\n",
    "# Treinando com todo o conjunto de dados\n",
    "history = final_model.fit(\n",
    "        data_train_sclr, train_targets,\n",
    "        epochs=10,\n",
    "        batch_size=256,\n",
    "        verbose=1, # Setting the model fitting visualization\n",
    "        callbacks=None\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualizing the model - Netron:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_model.save('models/model_2.keras')\n",
    "\n",
    "# No terminal, execute: netron model.h5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import load_model\n",
    "\n",
    "final_model = load_model('models/model_2.keras')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Serving 'models/model_2.keras' at http://localhost:8080\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "('localhost', 8080)"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import netron\n",
    "\n",
    "# Substitua o caminho pelo seu arquivo de modelo salvo\n",
    "netron.start('models/model_2.keras')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Measuring performance with Validation Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 5917,
     "status": "ok",
     "timestamp": 1734382064684,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "WSpXADIe9aJc",
    "outputId": "728e29b5-9f5e-4d3a-f5bc-d405cd912df1"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m16115/16115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m13s\u001b[0m 800us/step\n",
      "(515668, 1)\n"
     ]
    }
   ],
   "source": [
    "val_predictions_prob = model.predict(data_validation_sclr)\n",
    "print(val_predictions_prob.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.45922184],\n",
       "       [0.45950538],\n",
       "       [0.9716173 ],\n",
       "       ...,\n",
       "       [0.96528023],\n",
       "       [0.31896484],\n",
       "       [0.49277964]], dtype=float32)"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "val_predictions_prob"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 178
    },
    "executionInfo": {
     "elapsed": 8,
     "status": "ok",
     "timestamp": 1734382070983,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "_ZWuTq7tCNkv",
    "outputId": "e7cecfc2-2a94-4c49-83ec-21bd3335e61f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0\n",
       "0    313613\n",
       "1    202055\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "threshold = 0.5\n",
    "val_predictions = (val_predictions_prob > threshold).astype(int).flatten()\n",
    "pd.DataFrame(val_predictions).value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 178
    },
    "executionInfo": {
     "elapsed": 357,
     "status": "ok",
     "timestamp": 1734382074729,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "3XV36kCXFE4E",
    "outputId": "39e7f397-48b4-482c-e8bc-fd86aa03c30d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "is_fraud\n",
       "1           257896\n",
       "0           257772\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(validation_targets).value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "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>Actual</th>\n",
       "      <th>Prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1697733</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>104627</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1969554</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>636899</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>893306</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1479935</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>750406</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1828680</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>129504</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>582520</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>515668 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         Actual  Prediction\n",
       "1697733       1           0\n",
       "104627        0           0\n",
       "1969554       1           1\n",
       "636899        0           0\n",
       "893306        0           0\n",
       "...         ...         ...\n",
       "1479935       1           1\n",
       "750406        0           0\n",
       "1828680       1           1\n",
       "129504        0           0\n",
       "582520        0           0\n",
       "\n",
       "[515668 rows x 2 columns]"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "output_data_frame = pd.DataFrame({'Actual': validation_targets, 'Prediction': val_predictions.flatten()})\n",
    "output_data_frame"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Validation Dataset Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {
    "executionInfo": {
     "elapsed": 208,
     "status": "ok",
     "timestamp": 1734382080703,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "YJ0_5CdcCQUh"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.97      0.88    257772\n",
      "           1       0.97      0.76      0.85    257896\n",
      "\n",
      "    accuracy                           0.87    515668\n",
      "   macro avg       0.88      0.87      0.86    515668\n",
      "weighted avg       0.88      0.87      0.86    515668\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "print(classification_report(validation_targets, val_predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "[[250901   6871]\n",
      " [ 62712 195184]]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "print(\"Confusion Matrix:\")\n",
    "print(confusion_matrix(validation_targets, val_predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8650624044928132\n",
      "Precision: 0.9659944074633144\n",
      "Recall: 0.7568322114340664\n",
      "F1 Score: 0.8487164937134608\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score\n",
    "\n",
    "accuracy = accuracy_score(validation_targets, val_predictions)\n",
    "precision = precision_score(validation_targets, val_predictions)\n",
    "recall = recall_score(validation_targets, val_predictions)\n",
    "f1 = f1_score(validation_targets, val_predictions)\n",
    "\n",
    "print(f\"Accuracy: {accuracy}\")\n",
    "print(f\"Precision: {precision}\")\n",
    "print(f\"Recall: {recall}\")\n",
    "print(f\"F1 Score: {f1}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC-AUC Score: 0.8650884363037532\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "\n",
    "roc_auc = roc_auc_score(validation_targets, val_predictions)\n",
    "print(f\"ROC-AUC Score: {roc_auc}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "\n",
    "fpr, tpr, thresholds = roc_curve(validation_targets, val_predictions)\n",
    "\n",
    "plt.plot(fpr, tpr, marker='.')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC Curve')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import precision_recall_curve\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "precision, recall, thresholds = precision_recall_curve(validation_targets, val_predictions)\n",
    "\n",
    "plt.plot(recall, precision, marker='.')\n",
    "plt.xlabel('Recall')\n",
    "plt.ylabel('Precision')\n",
    "plt.title('Precision-Recall Curve')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "h8lyFSjZIpua"
   },
   "source": [
    "## Making predictions with Test Data and Validating"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>trans_date_trans_time</th>\n",
       "      <th>cc_num</th>\n",
       "      <th>merchant</th>\n",
       "      <th>category</th>\n",
       "      <th>amt</th>\n",
       "      <th>first</th>\n",
       "      <th>last</th>\n",
       "      <th>gender</th>\n",
       "      <th>street</th>\n",
       "      <th>...</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "      <th>city_pop</th>\n",
       "      <th>job</th>\n",
       "      <th>dob</th>\n",
       "      <th>trans_num</th>\n",
       "      <th>unix_time</th>\n",
       "      <th>merch_lat</th>\n",
       "      <th>merch_long</th>\n",
       "      <th>is_fraud</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>2020-06-21 12:14:25</td>\n",
       "      <td>2291163933867244</td>\n",
       "      <td>fraud_Kirlin and Sons</td>\n",
       "      <td>personal_care</td>\n",
       "      <td>2.86</td>\n",
       "      <td>Jeff</td>\n",
       "      <td>Elliott</td>\n",
       "      <td>M</td>\n",
       "      <td>351 Darlene Green</td>\n",
       "      <td>...</td>\n",
       "      <td>33.9659</td>\n",
       "      <td>-80.9355</td>\n",
       "      <td>333497</td>\n",
       "      <td>Mechanical engineer</td>\n",
       "      <td>1968-03-19</td>\n",
       "      <td>2da90c7d74bd46a0caf3777415b3ebd3</td>\n",
       "      <td>1371816865</td>\n",
       "      <td>33.986391</td>\n",
       "      <td>-81.200714</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-06-21 12:14:33</td>\n",
       "      <td>3573030041201292</td>\n",
       "      <td>fraud_Sporer-Keebler</td>\n",
       "      <td>personal_care</td>\n",
       "      <td>29.84</td>\n",
       "      <td>Joanne</td>\n",
       "      <td>Williams</td>\n",
       "      <td>F</td>\n",
       "      <td>3638 Marsh Union</td>\n",
       "      <td>...</td>\n",
       "      <td>40.3207</td>\n",
       "      <td>-110.4360</td>\n",
       "      <td>302</td>\n",
       "      <td>Sales professional, IT</td>\n",
       "      <td>1990-01-17</td>\n",
       "      <td>324cc204407e99f51b0d6ca0055005e7</td>\n",
       "      <td>1371816873</td>\n",
       "      <td>39.450498</td>\n",
       "      <td>-109.960431</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>2020-06-21 12:14:53</td>\n",
       "      <td>3598215285024754</td>\n",
       "      <td>fraud_Swaniawski, Nitzsche and Welch</td>\n",
       "      <td>health_fitness</td>\n",
       "      <td>41.28</td>\n",
       "      <td>Ashley</td>\n",
       "      <td>Lopez</td>\n",
       "      <td>F</td>\n",
       "      <td>9333 Valentine Point</td>\n",
       "      <td>...</td>\n",
       "      <td>40.6729</td>\n",
       "      <td>-73.5365</td>\n",
       "      <td>34496</td>\n",
       "      <td>Librarian, public</td>\n",
       "      <td>1970-10-21</td>\n",
       "      <td>c81755dbbbea9d5c77f094348a7579be</td>\n",
       "      <td>1371816893</td>\n",
       "      <td>40.495810</td>\n",
       "      <td>-74.196111</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>2020-06-21 12:15:15</td>\n",
       "      <td>3591919803438423</td>\n",
       "      <td>fraud_Haley Group</td>\n",
       "      <td>misc_pos</td>\n",
       "      <td>60.05</td>\n",
       "      <td>Brian</td>\n",
       "      <td>Williams</td>\n",
       "      <td>M</td>\n",
       "      <td>32941 Krystal Mill Apt. 552</td>\n",
       "      <td>...</td>\n",
       "      <td>28.5697</td>\n",
       "      <td>-80.8191</td>\n",
       "      <td>54767</td>\n",
       "      <td>Set designer</td>\n",
       "      <td>1987-07-25</td>\n",
       "      <td>2159175b9efe66dc301f149d3d5abf8c</td>\n",
       "      <td>1371816915</td>\n",
       "      <td>28.812398</td>\n",
       "      <td>-80.883061</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>2020-06-21 12:15:17</td>\n",
       "      <td>3526826139003047</td>\n",
       "      <td>fraud_Johnston-Casper</td>\n",
       "      <td>travel</td>\n",
       "      <td>3.19</td>\n",
       "      <td>Nathan</td>\n",
       "      <td>Massey</td>\n",
       "      <td>M</td>\n",
       "      <td>5783 Evan Roads Apt. 465</td>\n",
       "      <td>...</td>\n",
       "      <td>44.2529</td>\n",
       "      <td>-85.0170</td>\n",
       "      <td>1126</td>\n",
       "      <td>Furniture designer</td>\n",
       "      <td>1955-07-06</td>\n",
       "      <td>57ff021bd3f328f8738bb535c302a31b</td>\n",
       "      <td>1371816917</td>\n",
       "      <td>44.959148</td>\n",
       "      <td>-85.884734</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0 trans_date_trans_time            cc_num  \\\n",
       "0           0   2020-06-21 12:14:25  2291163933867244   \n",
       "1           1   2020-06-21 12:14:33  3573030041201292   \n",
       "2           2   2020-06-21 12:14:53  3598215285024754   \n",
       "3           3   2020-06-21 12:15:15  3591919803438423   \n",
       "4           4   2020-06-21 12:15:17  3526826139003047   \n",
       "\n",
       "                               merchant        category    amt   first  \\\n",
       "0                 fraud_Kirlin and Sons   personal_care   2.86    Jeff   \n",
       "1                  fraud_Sporer-Keebler   personal_care  29.84  Joanne   \n",
       "2  fraud_Swaniawski, Nitzsche and Welch  health_fitness  41.28  Ashley   \n",
       "3                     fraud_Haley Group        misc_pos  60.05   Brian   \n",
       "4                 fraud_Johnston-Casper          travel   3.19  Nathan   \n",
       "\n",
       "       last gender                       street  ...      lat      long  \\\n",
       "0   Elliott      M            351 Darlene Green  ...  33.9659  -80.9355   \n",
       "1  Williams      F             3638 Marsh Union  ...  40.3207 -110.4360   \n",
       "2     Lopez      F         9333 Valentine Point  ...  40.6729  -73.5365   \n",
       "3  Williams      M  32941 Krystal Mill Apt. 552  ...  28.5697  -80.8191   \n",
       "4    Massey      M     5783 Evan Roads Apt. 465  ...  44.2529  -85.0170   \n",
       "\n",
       "   city_pop                     job         dob  \\\n",
       "0    333497     Mechanical engineer  1968-03-19   \n",
       "1       302  Sales professional, IT  1990-01-17   \n",
       "2     34496       Librarian, public  1970-10-21   \n",
       "3     54767            Set designer  1987-07-25   \n",
       "4      1126      Furniture designer  1955-07-06   \n",
       "\n",
       "                          trans_num   unix_time  merch_lat  merch_long  \\\n",
       "0  2da90c7d74bd46a0caf3777415b3ebd3  1371816865  33.986391  -81.200714   \n",
       "1  324cc204407e99f51b0d6ca0055005e7  1371816873  39.450498 -109.960431   \n",
       "2  c81755dbbbea9d5c77f094348a7579be  1371816893  40.495810  -74.196111   \n",
       "3  2159175b9efe66dc301f149d3d5abf8c  1371816915  28.812398  -80.883061   \n",
       "4  57ff021bd3f328f8738bb535c302a31b  1371816917  44.959148  -85.884734   \n",
       "\n",
       "   is_fraud  \n",
       "0         0  \n",
       "1         0  \n",
       "2         0  \n",
       "3         0  \n",
       "4         0  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 132,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_data = pd.read_csv(\"data/fraudTest.csv\")\n",
    "test_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "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>cc_num</th>\n",
       "      <th>amt</th>\n",
       "      <th>zip</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "      <th>city_pop</th>\n",
       "      <th>unix_time</th>\n",
       "      <th>merch_lat</th>\n",
       "      <th>merch_long</th>\n",
       "      <th>is_fraud</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2291163933867244</td>\n",
       "      <td>2.86</td>\n",
       "      <td>29209</td>\n",
       "      <td>33.9659</td>\n",
       "      <td>-80.9355</td>\n",
       "      <td>333497</td>\n",
       "      <td>1371816865</td>\n",
       "      <td>33.986391</td>\n",
       "      <td>-81.200714</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3573030041201292</td>\n",
       "      <td>29.84</td>\n",
       "      <td>84002</td>\n",
       "      <td>40.3207</td>\n",
       "      <td>-110.4360</td>\n",
       "      <td>302</td>\n",
       "      <td>1371816873</td>\n",
       "      <td>39.450498</td>\n",
       "      <td>-109.960431</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3598215285024754</td>\n",
       "      <td>41.28</td>\n",
       "      <td>11710</td>\n",
       "      <td>40.6729</td>\n",
       "      <td>-73.5365</td>\n",
       "      <td>34496</td>\n",
       "      <td>1371816893</td>\n",
       "      <td>40.495810</td>\n",
       "      <td>-74.196111</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3591919803438423</td>\n",
       "      <td>60.05</td>\n",
       "      <td>32780</td>\n",
       "      <td>28.5697</td>\n",
       "      <td>-80.8191</td>\n",
       "      <td>54767</td>\n",
       "      <td>1371816915</td>\n",
       "      <td>28.812398</td>\n",
       "      <td>-80.883061</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3526826139003047</td>\n",
       "      <td>3.19</td>\n",
       "      <td>49632</td>\n",
       "      <td>44.2529</td>\n",
       "      <td>-85.0170</td>\n",
       "      <td>1126</td>\n",
       "      <td>1371816917</td>\n",
       "      <td>44.959148</td>\n",
       "      <td>-85.884734</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>555714</th>\n",
       "      <td>30560609640617</td>\n",
       "      <td>43.77</td>\n",
       "      <td>63453</td>\n",
       "      <td>40.4931</td>\n",
       "      <td>-91.8912</td>\n",
       "      <td>519</td>\n",
       "      <td>1388534347</td>\n",
       "      <td>39.946837</td>\n",
       "      <td>-91.333331</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>555715</th>\n",
       "      <td>3556613125071656</td>\n",
       "      <td>111.84</td>\n",
       "      <td>77566</td>\n",
       "      <td>29.0393</td>\n",
       "      <td>-95.4401</td>\n",
       "      <td>28739</td>\n",
       "      <td>1388534349</td>\n",
       "      <td>29.661049</td>\n",
       "      <td>-96.186633</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>555716</th>\n",
       "      <td>6011724471098086</td>\n",
       "      <td>86.88</td>\n",
       "      <td>99323</td>\n",
       "      <td>46.1966</td>\n",
       "      <td>-118.9017</td>\n",
       "      <td>3684</td>\n",
       "      <td>1388534355</td>\n",
       "      <td>46.658340</td>\n",
       "      <td>-119.715054</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>555717</th>\n",
       "      <td>4079773899158</td>\n",
       "      <td>7.99</td>\n",
       "      <td>83643</td>\n",
       "      <td>44.6255</td>\n",
       "      <td>-116.4493</td>\n",
       "      <td>129</td>\n",
       "      <td>1388534364</td>\n",
       "      <td>44.470525</td>\n",
       "      <td>-117.080888</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>555718</th>\n",
       "      <td>4170689372027579</td>\n",
       "      <td>38.13</td>\n",
       "      <td>73034</td>\n",
       "      <td>35.6665</td>\n",
       "      <td>-97.4798</td>\n",
       "      <td>116001</td>\n",
       "      <td>1388534374</td>\n",
       "      <td>36.210097</td>\n",
       "      <td>-97.036372</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>555719 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                  cc_num     amt    zip      lat      long  city_pop  \\\n",
       "0       2291163933867244    2.86  29209  33.9659  -80.9355    333497   \n",
       "1       3573030041201292   29.84  84002  40.3207 -110.4360       302   \n",
       "2       3598215285024754   41.28  11710  40.6729  -73.5365     34496   \n",
       "3       3591919803438423   60.05  32780  28.5697  -80.8191     54767   \n",
       "4       3526826139003047    3.19  49632  44.2529  -85.0170      1126   \n",
       "...                  ...     ...    ...      ...       ...       ...   \n",
       "555714    30560609640617   43.77  63453  40.4931  -91.8912       519   \n",
       "555715  3556613125071656  111.84  77566  29.0393  -95.4401     28739   \n",
       "555716  6011724471098086   86.88  99323  46.1966 -118.9017      3684   \n",
       "555717     4079773899158    7.99  83643  44.6255 -116.4493       129   \n",
       "555718  4170689372027579   38.13  73034  35.6665  -97.4798    116001   \n",
       "\n",
       "         unix_time  merch_lat  merch_long  is_fraud  \n",
       "0       1371816865  33.986391  -81.200714         0  \n",
       "1       1371816873  39.450498 -109.960431         0  \n",
       "2       1371816893  40.495810  -74.196111         0  \n",
       "3       1371816915  28.812398  -80.883061         0  \n",
       "4       1371816917  44.959148  -85.884734         0  \n",
       "...            ...        ...         ...       ...  \n",
       "555714  1388534347  39.946837  -91.333331         0  \n",
       "555715  1388534349  29.661049  -96.186633         0  \n",
       "555716  1388534355  46.658340 -119.715054         0  \n",
       "555717  1388534364  44.470525 -117.080888         0  \n",
       "555718  1388534374  36.210097  -97.036372         0  \n",
       "\n",
       "[555719 rows x 10 columns]"
      ]
     },
     "execution_count": 133,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_data = test_data[num_columns]\n",
    "test_data.drop(\"Unnamed: 0\",axis=1,inplace=True)\n",
    "test_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [],
   "source": [
    "features_test = test_data.iloc[:, :-1]\n",
    "targets_test = test_data.iloc[:,-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_test_sclr = pd.DataFrame(\n",
    "    scaler.transform(features_test),\n",
    "    columns=features_test.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 4174,
     "status": "ok",
     "timestamp": 1734382091969,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "eOm0wcOsCjb_",
    "outputId": "fdf4e5ff-a133-4771-8d84-5747a6ef0cf9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m17367/17367\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m41s\u001b[0m 2ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[0.07176076],\n",
       "       [0.20699978],\n",
       "       [0.0467512 ],\n",
       "       ...,\n",
       "       [0.05955727],\n",
       "       [0.2790884 ],\n",
       "       [0.05694652]], dtype=float32)"
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_predictions_prob = final_model.predict(data_test_sclr)\n",
    "test_predictions_prob"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(555719, 1)"
      ]
     },
     "execution_count": 137,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_predictions_prob.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 218,
     "status": "ok",
     "timestamp": 1734382093532,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "w34jmruKJtJk",
    "outputId": "beb5d6b9-7127-46ba-af80-69f9867a0035"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.07176076, 0.20699978, 0.0467512 , ..., 0.05955727, 0.2790884 ,\n",
       "       0.05694652], dtype=float32)"
      ]
     },
     "execution_count": 138,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_predictions_prob.reshape(555719)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 178
    },
    "executionInfo": {
     "elapsed": 3,
     "status": "ok",
     "timestamp": 1734382094565,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "gXHpkp7AHYoG",
    "outputId": "df45fcc5-a154-4c04-d588-11074db9f47b"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0\n",
       "0    509433\n",
       "1     46286\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 139,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "threshold = 0.5\n",
    "test_predictions = (test_predictions_prob > threshold).astype(int).flatten()\n",
    "pd.DataFrame(test_predictions).value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 178
    },
    "executionInfo": {
     "elapsed": 245,
     "status": "ok",
     "timestamp": 1734382099799,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "NZJhvot0Hgxg",
    "outputId": "d64dce9d-db91-4e20-a62c-d438e5c0113e"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "is_fraud\n",
       "0           553574\n",
       "1             2145\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 140,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(targets_test).value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 424
    },
    "executionInfo": {
     "elapsed": 272,
     "status": "ok",
     "timestamp": 1734382104233,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "rjLPakapH0K8",
    "outputId": "2172adca-4b12-4411-fb5f-74b11d62c058"
   },
   "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>Actual</th>\n",
       "      <th>Prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>555714</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>555715</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>555716</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>555717</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>555718</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>555719 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        Actual  Prediction\n",
       "0            0           0\n",
       "1            0           0\n",
       "2            0           0\n",
       "3            0           0\n",
       "4            0           0\n",
       "...        ...         ...\n",
       "555714       0           0\n",
       "555715       0           0\n",
       "555716       0           0\n",
       "555717       0           0\n",
       "555718       0           0\n",
       "\n",
       "[555719 rows x 2 columns]"
      ]
     },
     "execution_count": 141,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "output_data_frame = pd.DataFrame({'Actual': targets_test, 'Prediction': test_predictions.flatten()})\n",
    "output_data_frame"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test Dataset Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 932,
     "status": "ok",
     "timestamp": 1734382487875,
     "user": {
      "displayName": "Danrlei Diegues",
      "userId": "13520636167171764574"
     },
     "user_tz": 180
    },
    "id": "qCiKh1EKI7nq",
    "outputId": "91365573-7b59-47f6-908b-ffe674af1337"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.92      0.96    553574\n",
      "           1       0.04      0.80      0.07      2145\n",
      "\n",
      "    accuracy                           0.92    555719\n",
      "   macro avg       0.52      0.86      0.51    555719\n",
      "weighted avg       1.00      0.92      0.95    555719\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "print(classification_report(targets_test, test_predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      "[[509009  44565]\n",
      " [   424   1721]]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "print(\"Confusion Matrix:\")\n",
    "print(confusion_matrix(targets_test, test_predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9190436173677704\n",
      "Precision: 0.03718186924772069\n",
      "Recall: 0.8023310023310023\n",
      "F1 Score: 0.07107018232124053\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score\n",
    "\n",
    "accuracy = accuracy_score(targets_test, test_predictions)\n",
    "precision = precision_score(targets_test, test_predictions)\n",
    "recall = recall_score(targets_test, test_predictions)\n",
    "f1 = f1_score(targets_test, test_predictions)\n",
    "\n",
    "print(f\"Accuracy: {accuracy}\")\n",
    "print(f\"Precision: {precision}\")\n",
    "print(f\"Recall: {recall}\")\n",
    "print(f\"F1 Score: {f1}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC-AUC Score: 0.8609134300783474\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "\n",
    "roc_auc = roc_auc_score(targets_test, test_predictions)\n",
    "print(f\"ROC-AUC Score: {roc_auc}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjcAAAHHCAYAAABDUnkqAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8ekN5oAAAACXBIWXMAAA9hAAAPYQGoP6dpAABNAklEQVR4nO3deXhTVf4G8DdJm6T7QulKoCyiIAiyTkFkkGpxQRlUqjBQUHEDZERGdgoulBFBHEEZUayAyjaojCD8BEUBURSogGy2tJSlK4Wme9rk/P6oCQ1pSlKS3DZ5P8+TR3p7b/LNdcnryfecIxNCCBARERG5CbnUBRARERE5EsMNERERuRWGGyIiInIrDDdERETkVhhuiIiIyK0w3BAREZFbYbghIiIit8JwQ0RERG6F4YaIiIjcCsMNERERuRWGGyJqUGpqKmQymenh5eWFmJgYjB07FhcuXKj3GiEE1qxZgzvvvBPBwcHw9fVF165d8corr6CsrMzqa33++ee49957ERYWBqVSiejoaIwYMQLffvutTbVWVlbirbfeQt++fREUFAS1Wo2OHTti4sSJOH36dKPePxE1PzLuLUVEDUlNTcW4cePwyiuvoG3btqisrMRPP/2E1NRUxMbG4tixY1Cr1abz9Xo9Ro4ciQ0bNmDAgAEYPnw4fH19sWfPHnz66afo3Lkzdu7ciYiICNM1Qgg88cQTSE1Nxe23345HHnkEkZGRyMnJweeff46DBw9i37596Nevn9U6CwsLMWTIEBw8eBAPPPAA4uPj4e/vj1OnTmHdunXIzc2FTqdz6r0ioiZCEBE14KOPPhIAxC+//GJ2fNq0aQKAWL9+vdnxBQsWCABi6tSpFs+1ZcsWIZfLxZAhQ8yOL1q0SAAQ//jHP4TBYLC4bvXq1eLnn39usM77779fyOVysWnTJovfVVZWipdeeqnB621VXV0tqqqqHPJcROQcDDdE1CBr4earr74SAMSCBQtMx8rLy0VISIjo2LGjqK6urvf5xo0bJwCI/fv3m64JDQ0Vt9xyi6ipqWlUjT/99JMAIMaPH2/T+QMHDhQDBw60OJ6UlCTatGlj+jkzM1MAEIsWLRJvvfWWaNeunZDL5eKnn34SCoVCzJs3z+I5Tp48KQCId955x3Ts8uXLYvLkyaJVq1ZCqVSK9u3bi4ULFwq9Xm/3eyWi62PPDRE1SlZWFgAgJCTEdGzv3r24fPkyRo4cCS8vr3qvGzNmDADgq6++Ml1TVFSEkSNHQqFQNKqWLVu2AABGjx7dqOuv56OPPsI777yDp59+GosXL0ZUVBQGDhyIDRs2WJy7fv16KBQKPProowCA8vJyDBw4EGvXrsWYMWPw73//G/3798eMGTMwZcoUp9RL5Onq/68PEdE1iouLUVhYiMrKSvz888+YP38+VCoVHnjgAdM5x48fBwB069bN6vMYf3fixAmzv3bt2rXRtTniORpy/vx5pKeno2XLlqZjiYmJeOaZZ3Ds2DF06dLFdHz9+vUYOHCgqadoyZIlyMjIwOHDh3HTTTcBAJ555hlER0dj0aJFeOmll6DRaJxSN5Gn4sgNEdkkPj4eLVu2hEajwSOPPAI/Pz9s2bIFrVq1Mp1TUlICAAgICLD6PMbfabVas782dM31OOI5GvLwww+bBRsAGD58OLy8vLB+/XrTsWPHjuH48eNITEw0Hdu4cSMGDBiAkJAQFBYWmh7x8fHQ6/X44YcfnFIzkSfjyA0R2WT58uXo2LEjiouLsWrVKvzwww9QqVRm5xjDhTHk1OfaABQYGHjda66n7nMEBwc3+nmsadu2rcWxsLAwDB48GBs2bMCrr74KoHbUxsvLC8OHDzed98cff+DIkSMW4cgoPz/f4fUSeTqGGyKySZ8+fdCrVy8AwLBhw3DHHXdg5MiROHXqFPz9/QEAnTp1AgAcOXIEw4YNq/d5jhw5AgDo3LkzAOCWW24BABw9etTqNddT9zkGDBhw3fNlMhlEPatg6PX6es/38fGp9/hjjz2GcePGIS0tDd27d8eGDRswePBghIWFmc4xGAy4++678fLLL9f7HB07drxuvURkH34tRUR2UygUSElJwcWLF7Fs2TLT8TvuuAPBwcH49NNPrQaF1atXA4CpV+eOO+5ASEgIPvvsM6vXXM/QoUMBAGvXrrXp/JCQEFy5csXi+NmzZ+163WHDhkGpVGL9+vVIS0vD6dOn8dhjj5md0759e5SWliI+Pr7eR+vWre16TSK6PoYbImqUv/71r+jTpw+WLl2KyspKAICvry+mTp2KU6dOYdasWRbXbN26FampqUhISMBf/vIX0zXTpk3DiRMnMG3atHpHVNauXYsDBw5YrSUuLg5DhgzBBx98gC+++MLi9zqdDlOnTjX93L59e5w8eRIFBQWmY7/99hv27dtn8/sHgODgYCQkJGDDhg1Yt24dlEqlxejTiBEjsH//fuzYscPi+itXrqCmpsau1ySi6+MKxUTUIOMKxb/88ovpaymjTZs24dFHH8V7772HZ599FkDtVzuJiYn473//izvvvBMPP/wwfHx8sHfvXqxduxadOnXCrl27zFYoNhgMGDt2LNasWYMePXqYVijOzc3FF198gQMHDuDHH39EXFyc1ToLCgpwzz334LfffsPQoUMxePBg+Pn54Y8//sC6deuQk5ODqqoqALWzq7p06YJu3brhySefRH5+PlasWIGIiAhotVrTNPesrCy0bdsWixYtMgtHdX3yySf4+9//joCAAPz1r381TUs3Ki8vx4ABA3DkyBGMHTsWPXv2RFlZGY4ePYpNmzYhKyvL7GssInIAaZfZIaKmztoifkIIodfrRfv27UX79u3NFuDT6/Xio48+Ev379xeBgYFCrVaLW2+9VcyfP1+UlpZafa1NmzaJe+65R4SGhgovLy8RFRUlEhMTxe7du22qtby8XLz55puid+/ewt/fXyiVSnHTTTeJSZMmifT0dLNz165dK9q1ayeUSqXo3r272LFjR4OL+Fmj1WqFj4+PACDWrl1b7zklJSVixowZokOHDkKpVIqwsDDRr18/8eabbwqdTmfTeyMi23HkhoiIiNwKe26IiIjIrTDcEBERkVthuCEiIiK3wnBDREREboXhhoiIiNwKww0RERG5FY/bW8pgMODixYsICAiATCaTuhwiIiKygRACJSUliI6Ohlze8NiMx4WbixcvQqPRSF0GERERNcK5c+fQqlWrBs/xuHATEBAAoPbmBAYGSlwNERER2UKr1UKj0Zg+xxviceHG+FVUYGAgww0REVEzY0tLCRuKiYiIyK0w3BAREZFbYbghIiIit8JwQ0RERG6F4YaIiIjcCsMNERERuRWGGyIiInIrDDdERETkVhhuiIiIyK0w3BAREZFbkTTc/PDDDxg6dCiio6Mhk8nwxRdfXPea3bt3o0ePHlCpVOjQoQNSU1OdXicRERE1H5KGm7KyMnTr1g3Lly+36fzMzEzcf//9GDRoENLS0vCPf/wDTz31FHbs2OHkSomIiMgWOcUV+DGjEDnFFZLVIOnGmffeey/uvfdem89fsWIF2rZti8WLFwMAOnXqhL179+Ktt95CQkKCs8okIiIiG6w7kI2Znx+FQQByGZAyvCsSe7d2eR3Nalfw/fv3Iz4+3uxYQkIC/vGPf1i9pqqqClVVVaaftVqts8ojIiLyCHqDwPnL5UjPLzU9judo8fvFq5+xBgHM3HwMd3ZsiaggH5fW16zCTW5uLiIiIsyORUREQKvVoqKiAj4+ljcvJSUF8+fPd1WJREREbqOqRo/MwjKzEJOeX4rMwjJU1Riue71eCGQVljPcONqMGTMwZcoU089arRYajUbCioiIiJqWksrqq+GloBQZf/45u6gcBlH/NSovOdq19EeHcH90aOmPFv5KzP3ymNn5CpkMsWG+rnkTdTSrcBMZGYm8vDyzY3l5eQgMDKx31AYAVCoVVCqVK8ojIiJqsoQQKCitQnr+1fCSXlD71zxtldXrAtVetQGm7qNlAGJCfKCQy8zO9VbIMHPzMeiFgEImw4LhXVw+agM0s3ATFxeHbdu2mR375ptvEBcXJ1FFRERETYvBIHD+cgXSC0osvk7SVtZYvS48QHVNgKn9a8sAFWQymdXr6krs3Rp3dmyJrMJyxIb5ShJsAInDTWlpKdLT000/Z2ZmIi0tDaGhoWjdujVmzJiBCxcuYPXq1QCAZ599FsuWLcPLL7+MJ554At9++y02bNiArVu3SvUWiIiIJFFVo0dWYbnZ10np+aU4U1BqtR9GLgM0ob6m4NK+TpgJVHs7pK6oIB/JQo2RpOHm119/xaBBg0w/G3tjkpKSkJqaipycHGRnZ5t+37ZtW2zduhUvvvgi3n77bbRq1QoffPABp4ETEZHbKqmsRkaBeVNvRkFtP4zeSkOM0kuOdmF+teGl5dUA0zbMD2pvhYvfgevJhBBWWoXck1arRVBQEIqLixEYGCh1OURERBBCoLBUZ9HQm55filxtpdXrAlReZqMvxiCjCfW16Idp7uz5/G5WPTdERETNmcEgcOFKhXkvzJ9fJxVXVFu9rmWAymwExvgIt6MfxpMw3BARETmYrsaArEuW68OcKSxFZXX9/TAyGaAJ8TUbhTGOygT5OKYfxlMw3BARETVSaVWNxbTqjPxSnG2oH0YhR9swP/OG3pb+aNfSM/phXIHhhoiI6DoulVbhj2saetPzS5FTbL0fxt/YD3PN10maEB94KSTdt9rtMdwQERGhTj/MNQ29GQWluFxuvR8mzF+FDuF+dRp6A9Ah3B8RgeyHkQrDDREReRRdjQFn6/bDmNaHKUNFtb7ea2QyoFWIj2VTb8sABPmyH6apYbghIiK3VK6rQUZ+mdlKvX/klyL7UjlqrPTDeCtkpn6Yug297cL84aNkP0xzwXBDRETNWlGZzmJqdUZ+KS5cqbB6jZ9SYdHQ2yHcH61DfdkP4wYYboiIqMkTQuBicaX5Kr1/BpmiMp3V68L8lWhfz/owkYFq9sO4MYYbIiJqMqr1Bpy9VG42I8n453Jd/f0wwJ/9MPXMTAr2VbqwemoqGG6IiMjlynU1OHPNfknpBaXIKixrsB8mtoWfWXhp37L2wX4YqovhhoiInOZymc40G6nuo6F+GN8/+2HqNvQa+2G82Q9DNmC4ISKiGyKEQE7dfpg6K/VeaqAfJtRPaQowN9UJMVFB7IehG8NwQ0RENqnRG3C2qNyioTcjvxRlDfTDxAT71LtSb6gf+2HIORhuiIjITIVOj4wC84be9PxSZF0qQ7W+/n4YL7kMbVr4Wixw166lH/xU/Kgh1+I/cUREHupKueX6MMZ+GFF/hoGPtwLtw/0sRmHatPBjPww1GQw3RERuTAiBXG2lRUNvRkEpCkut98OE+HqbzUgy/jk6yAdyOfthqGljuCEicgM1egOyjf0wdRp6MwrKUFpVY/W66CC12Ywk44hMC3+VC6snciyGGyKiZqSyWm/qhcmoE2SyCsuh0xvqvUZh7Ie55quk9i392Q9Dbon/VBMRNUHF5dVmGz4aR2TOX7beD6P2ll/9CqmleT+M0ov9MOQ5GG6IiCQihECeturP8FJSZ7G7MhSWVlm9LtjX2yy8GKdZxwSzH4YIYLghInI6vUFc7YepMwpzJr8UJQ30w0QFqS0aejuE+6OFn5KL3BE1gOGGiMhBKqv1tfsl1e2JyS9FZmFZw/0wob5Xm3pbXh2N8Wc/DFGj8N8cIiI7FVdUWzT0pueX4tzlcqv9MCovucUITG0/jC9UXtz0kciRGG6IiOohhEB+SZXF+jDpBaUoKLHeDxPk423R0NshnP0wRK7EcENEHk1vEDh3zfowxkXuSiqt98NEBqotGno7hPsjzJ/9MERSY7ghIo9QWa1HZmGZ2QhMRn4pzhSWQVdTfz+MXAa0aeFn8XVS+5Z+CFB7u/gdEJGtGG6IyK1oK6vNd63+M8icKyqHoYF+mHb1rA8TG8Z+GKLmiOGGiJodIQQKjP0w1+xcnd9AP0yg2suiobdDywDEhPhAwX4YIrfBcENETZbeIHDhcoVppd4/8q5+naRtoB8mPEB1TYCp/WvLABX7YYg8AMMNEUmuqkaPrMJys36Y9PxSnCkoRVUD/TCaUF+zdWGMYSaQ/TBEHo3hhohcpqSyGhkFZWZfI2UUlOLspTKr/TBKLznahfmZzUjqEO6PtmF+UHuzH4aILDHcEJFDCSFQWKozm5FkDDK52kqr1wWovMxGX4xBRhPqy34YIrILww0RNYrBIHDhSoXFAnfp+aUorqi2el3LAJXFAncdwv0Rzn4YInIQhhsiapCuxoCsS2UWK/WeKSxFZXX9/TAyGaAJ8TUbhTGOygT5sB+GiJyL4YaIAAClVTVm68IY14k5W1QOvZWGGKVCjrZhfuYNvS390a4l+2GISDoMN0QeRAiBS2U6i4be9PxS5BRb74fxN/bDXPN1kibEB14KuQvfARHR9THcELkhUz/MNQ296QWluFJuvR8mzF+FDuF+dRp6A9Ah3B8RgeyHIaLmg+GGqBnT1Rhw9lJZPevDlKGiWl/vNTIZ0CrEx7Kpt2UAgnzZD0NEzR/DDVEzUFZVY/r6qG6QOXvJej+Mt0Jm6oep29DbLswfPkr2wxCR+2K4IWpCLpVa7peUkV+Kiw30w/gpFRYNvR3C/dE61Jf9METkkRhuiFzMYBC4WFxh0dCbnl+Kyw32wyjRvp71YSID1eyHISKqg+GGyEmq9df0w5hW7LXeDwP82Q9zzcyk9i39EeKndGH1RETNF8MN0Q0q19UgI7/MtHO18XH2UjlqrPTDeMlliA3zs2jqbd+S/TBERDeK4YbIRkXXrA9jnGZ94UqF1Wt8lQqzr5KMf27Twhfe7IchInIKhhuiOoQQuFhcab7I3Z9BpqhMZ/W6UD+l2Ywk4yMqUA05N30kInIphhvySLX9MOUWDb0ZBaUo11nvh4kJ9ql3pd5Q9sMQETUZDDfk1ip0eivrw5ShWm+9H6ZNC1+LBe7atfSDn4r/yhARNXX8LzW5hctlOrO1YYyPhvphfLwVaB9u2dTbpoUf+2GIiJoxhhtqNoQQyKnbD1Nn5+pLDfTDhPh6WzT0dgj3R3SQD/thiIjcEMMNNTk1egPOFpVbNPRm5JeirIF+mOggtXlD759BpoW/yoXVExGR1BhuSDLGfphre2KyGuiHURj7YepZH4b9MEREBDDckAtcKbdcH8bYDyPqzzBQe8uvfoXU0rwfRunFfhgiIrKO4YYcQgiBXG2lRUNvRkEZCkurrF4X7Ottvs3An2EmJpj9MERE1DgMN2SXGr0B2cZ+mDoNvRkFZSitqrF6XVSQ2qKht0O4P1r4KbnpIxERORTDDdWrslqPMwVlZgEmPb8UmYVl0OkN9V6jkMvQJtT3alNvy6ujMf7shyEiIheR/BNn+fLlWLRoEXJzc9GtWze888476NOnj9Xzly5divfeew/Z2dkICwvDI488gpSUFKjVahdW3TzkFFcgs7AMbcP8EBXkU+85xRXVZjOSjF8nnbtcbrUfRuUltxiBMe6XpPLipo9ERCQtScPN+vXrMWXKFKxYsQJ9+/bF0qVLkZCQgFOnTiE8PNzi/E8//RTTp0/HqlWr0K9fP5w+fRpjx46FTCbDkiVLJHgHTdf6X7IxY/NRGAQglwEz7u2EztGBFo29BSXW+2GCfLwtGno7hLMfhoiImjaZENb+/9z5+vbti969e2PZsmUAAIPBAI1Gg0mTJmH69OkW50+cOBEnTpzArl27TMdeeukl/Pzzz9i7d69Nr6nVahEUFITi4mIEBgY65o00MTnFFei/8FsYbPw7Gxmotmjo7RDujzB/9sMQEVHTYM/nt2QjNzqdDgcPHsSMGTNMx+RyOeLj47F///56r+nXrx/Wrl2LAwcOoE+fPjhz5gy2bduG0aNHW32dqqoqVFVdHZ3QarWOexNNVGZhWb3BJiJQha4xwdesD+OHALW364skIiJyEsnCTWFhIfR6PSIiIsyOR0RE4OTJk/VeM3LkSBQWFuKOO+6AEAI1NTV49tlnMXPmTKuvk5KSgvnz5zu09qaubZgf5DKYBRy5DPhiQn+rvTdERETuolmthrZ7924sWLAA7777Lg4dOoTNmzdj69atePXVV61eM2PGDBQXF5se586dc2HF0ogK8sEzA9ubflbIZEgZ3pXBhoiIPIJkIzdhYWFQKBTIy8szO56Xl4fIyMh6r5kzZw5Gjx6Np556CgDQtWtXlJWV4emnn8asWbMgl1tmNZVKBZXK8/YW6hjhDwDoFBWAVWN7M9gQEZHHkGzkRqlUomfPnmbNwQaDAbt27UJcXFy915SXl1sEGIWiduqxhH3RTVL2pQoAQNeYIAYbIiLyKJJOBZ8yZQqSkpLQq1cv9OnTB0uXLkVZWRnGjRsHABgzZgxiYmKQkpICABg6dCiWLFmC22+/HX379kV6ejrmzJmDoUOHmkIO1couKgcAtGnhJ3ElREREriVpuElMTERBQQHmzp2L3NxcdO/eHdu3bzc1GWdnZ5uN1MyePRsymQyzZ8/GhQsX0LJlSwwdOhSvv/66VG+hycouKgMAaEJ9Ja6EiIjItSRd50YKnrDODQD0XbATedoqfDmhP7ppgqUuh4iI6IbY8/ndrGZLkW0qq/XI09au7dOaIzdERORhGG7c0Lk/+20C1F4I9uUCfURE5FkYbtzQ2Uu14aZ1qC+3TyAiIo/DcOOGrs6U4ldSRETkeRhu3JAx3HCmFBEReSKGGzdkGrkJ5Ro3RETkeRhu3NDZS7Vr3HCmFBEReSKGGzdjMAicu1y79QJ7boiIyBMx3LiZvJJK6GoM8JLLEBWklrocIiIil2O4cTPZf04DjwnxgZeCf3uJiMjz8NPPzZwturrGDRERkSdiuHEz5xhuiIjIwzHcuJlshhsiIvJwDDdupu7WC0RERJ6I4cbNmL6W4jRwIiLyUAw3bqS0qgaXynQAOHJDRESei+HGjRingYf6KRGg9pa4GiIiImkw3LiR7KLabRe4YSYREXkyhhs3cnXDTIYbIiLyXAw3boQzpYiIiBhu3Eo2Z0oREREx3LgTLuBHRETEcOM2avQGXLhcAQBow5EbIiLyYAw3biKnuBI1BgGllxwRAWqpyyEiIpIMw42bMH4lpQnxgVwuk7gaIiIi6TDcuAnOlCIiIqrFcOMmTGvctPCTuBIiIiJpMdy4Ca5OTEREVIvhxk1wdWIiIqJaDDduwrhpJhfwIyIiT8dw4waulOugrawBAGhCGG6IiMizMdy4AeNXUuEBKvgoFRJXQ0REJC2GGzfAaeBERERX3VC4qaysdFQddAO4YSYREdFVdocbg8GAV199FTExMfD398eZM2cAAHPmzMGHH37o8ALp+rI5ckNERGRid7h57bXXkJqaijfeeANKpdJ0vEuXLvjggw8cWhzZ5uoCfgw3REREdoeb1atX4/3338eoUaOgUFxtXu3WrRtOnjzp0OLINqavpThyQ0REZH+4uXDhAjp06GBx3GAwoLq62iFFke10NQZcLK4AALQO5dYLREREdoebzp07Y8+ePRbHN23ahNtvv90hRZHtzl8uhxCAr1KBMH/l9S8gIiJyc172XjB37lwkJSXhwoULMBgM2Lx5M06dOoXVq1fjq6++ckaN1IC6X0nJZDKJqyEiIpKe3SM3Dz30EP73v/9h586d8PPzw9y5c3HixAn873//w9133+2MGqkBxnDDDTOJiIhq2T1yAwADBgzAN9984+haqBGM08C5YSYREVEtu0du2rVrh0uXLlkcv3LlCtq1a+eQosh2Z7mAHxERkRm7w01WVhb0er3F8aqqKly4cMEhRZHtznEaOBERkRmbv5basmWL6c87duxAUFCQ6We9Xo9du3YhNjbWocVRw4QQXOOGiIjoGjaHm2HDhgEAZDIZkpKSzH7n7e2N2NhYLF682KHFUcMKS3Uo1+khkwGtQhhuiIiIADvCjcFgAAC0bdsWv/zyC8LCwpxWFNnGOGoTHeQDpRc3eCciIgIaMVsqMzPTGXVQI2QXlQEANKE+EldCRETUdDRqKnhZWRm+//57ZGdnQ6fTmf3uhRdecEhhdH3Zl2q3XWjDbReIiIhM7A43hw8fxn333Yfy8nKUlZUhNDQUhYWF8PX1RXh4OMONC539c+SG08CJiIiusrtR48UXX8TQoUNx+fJl+Pj44KeffsLZs2fRs2dPvPnmm86okazgNHAiIiJLdoebtLQ0vPTSS5DL5VAoFKiqqoJGo8Ebb7yBmTNnOqNGsuLsJYYbIiKia9kdbry9vSGX114WHh6O7OxsAEBQUBDOnTvn2OrIqgqdHvklVQCANvxaioiIyMTunpvbb78dv/zyC2666SYMHDgQc+fORWFhIdasWYMuXbo4o0aqx7nLtaM2AWovBPl4S1wNERFR02H3yM2CBQsQFRUFAHj99dcREhKC5557DgUFBfjPf/7j8AKpfqYNM1v4QiaTSVwNERFR02H3yE2vXr1Mfw4PD8f27dsdWhDZ5iybiYmIiOrlsGVtDx06hAceeMDu65YvX47Y2Fio1Wr07dsXBw4caPD8K1euYMKECYiKioJKpULHjh2xbdu2xpbdbF2dKcU1boiIiOqyK9zs2LEDU6dOxcyZM3HmzBkAwMmTJzFs2DD07t3btEWDrdavX48pU6YgOTkZhw4dQrdu3ZCQkID8/Px6z9fpdLj77ruRlZWFTZs24dSpU1i5ciViYmLsel13cPbSn2vccOSGiIjIjM1fS3344YcYP348QkNDcfnyZXzwwQdYsmQJJk2ahMTERBw7dgydOnWy68WXLFmC8ePHY9y4cQCAFStWYOvWrVi1ahWmT59ucf6qVatQVFSEH3/8Ed7etU20nroTuXFfKc6UIiIiMmfzyM3bb7+Nf/3rXygsLMSGDRtQWFiId999F0ePHsWKFSvsDjY6nQ4HDx5EfHz81WLkcsTHx2P//v31XrNlyxbExcVhwoQJiIiIQJcuXbBgwQLo9Xqrr1NVVQWtVmv2aO4MBoFzl2u3XuDIDRERkTmbw01GRgYeffRRAMDw4cPh5eWFRYsWoVWrVo164cLCQuj1ekRERJgdj4iIQG5ubr3XnDlzBps2bYJer8e2bdswZ84cLF68GK+99prV10lJSUFQUJDpodFoGlVvU5JXUgldjQFechmigtRSl0NERNSk2BxuKioq4OtbO0ogk8mgUqlMU8JdxWAwIDw8HO+//z569uyJxMREzJo1CytWrLB6zYwZM1BcXGx6uMNCg8Zp4DEhPvBSOKwnnIiIyC3YNRX8gw8+gL+/PwCgpqYGqampCAsLMzvH1o0zw8LCoFAokJeXZ3Y8Ly8PkZGR9V4TFRUFb29vKBQK07FOnTohNzcXOp0OSqXS4hqVSgWVSmVTTc0Fp4ETERFZZ3O4ad26NVauXGn6OTIyEmvWrDE7RyaT2RxulEolevbsiV27dmHYsGEAakdmdu3ahYkTJ9Z7Tf/+/fHpp5/CYDCYtoA4ffo0oqKi6g027oobZhIREVlnc7jJyspy+ItPmTIFSUlJ6NWrF/r06YOlS5eirKzMNHtqzJgxiImJQUpKCgDgueeew7JlyzB58mRMmjQJf/zxBxYsWGBzoHIX3DCTiIjIOrtXKHakxMREFBQUYO7cucjNzUX37t2xfft2U5Nxdna2aYQGADQaDXbs2IEXX3wRt912G2JiYjB58mRMmzZNqrcgCU4DJyIisk4mhBBSF+FKWq0WQUFBKC4uRmBgoNTlNEqPV79BUZkOW1+4A7dGB0ldDhERkdPZ8/nNqTbNTEllNYrKdAD4tRQREVF9GG6aGeNXUqF+SgSovSWuhoiIqOlhuGlmOFOKiIioYY0KNxkZGZg9ezYef/xx0yaXX3/9NX7//XeHFkeWOFOKiIioYXaHm++//x5du3bFzz//jM2bN6O0tBQA8NtvvyE5OdnhBZI5zpQiIiJqmN3hZvr06XjttdfwzTffmC2cd9ddd+Gnn35yaHFkyRhuNBy5ISIiqpfd4ebo0aP429/+ZnE8PDwchYWFDimKrDON3DDcEBER1cvucBMcHIycnByL44cPH0ZMTIxDiqL61egNuHC5AgDQml9LERER1cvucPPYY49h2rRpyM3NhUwmg8FgwL59+zB16lSMGTPGGTXSn3KKK1FjEFB6yRERoJa6HCIioibJ7nCzYMEC3HLLLdBoNCgtLUXnzp1x5513ol+/fpg9e7YzaqQ/GWdKaUJ8IJfLJK6GiIioabJ7bymlUomVK1dizpw5OHbsGEpLS3H77bfjpptuckZ9VMfVmVJ+EldCRETUdNkdbvbu3Ys77rgDrVu3RuvWrZ1RE1mRzQX8iIiIrsvur6XuuusutG3bFjNnzsTx48edURNZkV1UBoDhhoiIqCF2h5uLFy/ipZdewvfff48uXbqge/fuWLRoEc6fP++M+qgOjtwQERFdn93hJiwsDBMnTsS+ffuQkZGBRx99FB9//DFiY2Nx1113OaNGAiCEuLr1AqeBExERWXVDG2e2bdsW06dPx8KFC9G1a1d8//33jqqLrlFcUY2SyhoAgCaE4YaIiMiaRoebffv24fnnn0dUVBRGjhyJLl26YOvWrY6sjeowjtqEB6jgo1RIXA0REVHTZfdsqRkzZmDdunW4ePEi7r77brz99tt46KGH4OvL0QRn4oaZREREtrE73Pzwww/45z//iREjRiAsLMwZNVE9uGEmERGRbewON/v27XNGHXQd2ZeMG2ZyAT8iIqKG2BRutmzZgnvvvRfe3t7YsmVLg+c++OCDDimMzJ01rnHTwkfiSoiIiJo2m8LNsGHDkJubi/DwcAwbNszqeTKZDHq93lG1UR3niv7cDZwjN0RERA2yKdwYDIZ6/0yuUVWjx8ViY7hhzw0REVFD7J4Kvnr1alRVVVkc1+l0WL16tUOKInMXLldACMBXqUCYv1LqcoiIiJo0u8PNuHHjUFxcbHG8pKQE48aNc0hRZO5snW0XZDKZxNUQERE1bXaHGyFEvR+w58+fR1BQkEOKInPnuKcUERGRzWyeCn777bdDJpNBJpNh8ODB8PK6eqler0dmZiaGDBnilCI9nWlPKYYbIiKi67I53BhnSaWlpSEhIQH+/v6m3ymVSsTGxuLhhx92eIHE1YmJiIjsYXO4SU5OBgDExsYiMTERarXaaUWRuXNcnZiIiMhmdq9QnJSU5Iw6yAohRJ2RG65xQ0REdD02hZvQ0FCcPn0aYWFhCAkJaXDGTlFRkcOKI6CwVIdynR5yGRATzNWJiYiIrsemcPPWW28hICDA9GdOR3ad7D+3XYgK8oHSy+7JbURERB7HpnBT96uosWPHOqsWqkc2p4ETERHZxe6hgEOHDuHo0aOmn7/88ksMGzYMM2fOhE6nc2hxxGngRERE9rI73DzzzDM4ffo0AODMmTNITEyEr68vNm7ciJdfftnhBXo608gNp4ETERHZxO5wc/r0aXTv3h0AsHHjRgwcOBCffvopUlNT8d///tfR9Xm8bI7cEBER2aVR2y8YdwbfuXMn7rvvPgCARqNBYWGhY6sjLuBHRERkJ7vDTa9evfDaa69hzZo1+P7773H//fcDADIzMxEREeHwAj1ZhU6P/JLaHdg5ckNERGQbu8PN0qVLcejQIUycOBGzZs1Chw4dAACbNm1Cv379HF6gJzt3uXbUJlDthWBfpcTVEBERNQ92r1B82223mc2WMlq0aBEUCoVDiqJapplS/EqKiIjIZnaHG6ODBw/ixIkTAIDOnTujR48eDiuKapn6bUK57QIREZGt7A43+fn5SExMxPfff4/g4GAAwJUrVzBo0CCsW7cOLVu2dHSNHiv7Uu3qxNwwk4iIyHZ299xMmjQJpaWl+P3331FUVISioiIcO3YMWq0WL7zwgjNq9FicKUVERGQ/u0dutm/fjp07d6JTp06mY507d8by5ctxzz33OLQ4T8etF4iIiOxn98iNwWCAt7e3xXFvb2/T+jd04wwGgXOXKwAw3BAREdnD7nBz1113YfLkybh48aLp2IULF/Diiy9i8ODBDi3Ok+WVVEJXY4CXXIaoILXU5RARETUbdoebZcuWQavVIjY2Fu3bt0f79u3Rtm1baLVavPPOO86o0SMZp4G3CvGBl8Luv01EREQey+6eG41Gg0OHDmHXrl2mqeCdOnVCfHy8w4vzZMZ+G86UIiIiso9d4Wb9+vXYsmULdDodBg8ejEmTJjmrLo9n3DCTM6WIiIjsY3O4ee+99zBhwgTcdNNN8PHxwebNm5GRkYFFixY5sz6PxZlSREREjWNzM8eyZcuQnJyMU6dOIS0tDR9//DHeffddZ9bm0c4y3BARETWKzeHmzJkzSEpKMv08cuRI1NTUICcnxymFebpzpnDDrReIiIjsYXO4qaqqgp/f1Q9auVwOpVKJiooKpxTmyUoqq1FUpgPATTOJiIjsZVdD8Zw5c+Dre/XDVqfT4fXXX0dQUJDp2JIlSxxXnYcy9tu08FPCX9XovU2JiIg8ks2fnHfeeSdOnTpldqxfv344c+aM6WeZTOa4yjyYcaYUp4ETERHZz+Zws3v3bieWQXVxw0wiIqLGaxJL3y5fvhyxsbFQq9Xo27cvDhw4YNN169atg0wmw7Bhw5xboItxphQREVHjSR5u1q9fjylTpiA5ORmHDh1Ct27dkJCQgPz8/Aavy8rKwtSpUzFgwAAXVeo65xhuiIiIGk3ycLNkyRKMHz8e48aNQ+fOnbFixQr4+vpi1apVVq/R6/UYNWoU5s+fj3bt2rmwWtcw7ivFcENERGQ/ScONTqfDwYMHzfalksvliI+Px/79+61e98orryA8PBxPPvmkK8p0qRq9AReu1E6vb9OCa9wQERHZS9J5xoWFhdDr9YiIiDA7HhERgZMnT9Z7zd69e/Hhhx8iLS3NpteoqqpCVVWV6WetVtvoel0hp7gSeoOA0kuO8ACV1OUQERE1O40audmzZw/+/ve/Iy4uDhcuXAAArFmzBnv37nVocdcqKSnB6NGjsXLlSoSFhdl0TUpKCoKCgkwPjUbj1BpvVN2vpORyTq0nIiKyl93h5r///S8SEhLg4+ODw4cPm0ZFiouLsWDBArueKywsDAqFAnl5eWbH8/LyEBkZaXF+RkYGsrKyMHToUHh5ecHLywurV6/Gli1b4OXlhYyMDItrZsyYgeLiYtPj3LlzdtXoatwwk4iI6MbYHW5ee+01rFixAitXroS3t7fpeP/+/XHo0CG7nkupVKJnz57YtWuX6ZjBYMCuXbsQFxdncf4tt9yCo0ePIi0tzfR48MEHMWjQIKSlpdU7KqNSqRAYGGj2aMrOFpUBYLghIiJqLLt7bk6dOoU777zT4nhQUBCuXLlidwFTpkxBUlISevXqhT59+mDp0qUoKyvDuHHjAABjxoxBTEwMUlJSoFar0aVLF7Prg4ODAcDieHPFaeBEREQ3xu5wExkZifT0dMTGxpod37t3b6OmZScmJqKgoABz585Fbm4uunfvju3bt5uajLOzsyGXSz5j3WWMPTdcnZiIiKhx7A4348ePx+TJk7Fq1SrIZDJcvHgR+/fvx9SpUzFnzpxGFTFx4kRMnDix3t9db9uH1NTURr1mUySEMO0rxZEbIiKixrE73EyfPh0GgwGDBw9GeXk57rzzTqhUKkydOhWTJk1yRo0e40p5NUqqagBw00wiIqLGsjvcyGQyzJo1C//85z+Rnp6O0tJSdO7cGf7+/s6oz6MYZ0pFBKqg9lZIXA0REVHz1OhF/JRKJTp37uzIWjweN8wkIiK6cXaHm0GDBkEms7643LfffntDBXmyqzOluO0CERFRY9kdbrp37272c3V1NdLS0nDs2DEkJSU5qi6PdPYS17ghIiK6UXaHm7feeqve4/PmzUNpaekNF+TJjD03nAZORETUeA5bQObvf/87Vq1a5ain80jGaeCcKUVERNR4Dgs3+/fvh1qtdtTTeZyqGj1ytJUAOHJDRER0I+z+Wmr48OFmPwshkJOTg19//bXRi/gRcP5yBYQAfJUKtPBTSl0OERFRs2V3uAkKCjL7WS6X4+abb8Yrr7yCe+65x2GFeZq6u4E3NBuNiIiIGmZXuNHr9Rg3bhy6du2KkJAQZ9XkkbhhJhERkWPY1XOjUChwzz33NGr3b2oYN8wkIiJyDLsbirt06YIzZ844oxaPls2RGyIiIoewO9y89tprmDp1Kr766ivk5ORAq9WaPahxTLuBt+DqxERERDfC5p6bV155BS+99BLuu+8+AMCDDz5o1vgqhIBMJoNer3d8lW5OCMGRGyIiIgexOdzMnz8fzz77LL777jtn1uORCkqrUFGth1wGxAT7SF0OERFRs2ZzuBFCAAAGDhzotGI8lXGmVFSQD5ReDltXkYiIyCPZ9UnK9VecwzhTil9JERER3Ti71rnp2LHjdQNOUVHRDRXkibhhJhERkePYFW7mz59vsUIx3ThumElEROQ4doWbxx57DOHh4c6qxWNx5IaIiMhxbO65Yb+N85zlNHAiIiKHsTncGGdLkWNV6PQoKKkCALQJ5QJ+REREN8rmr6UMBoMz6/BYxq+kAtVeCPL1lrgaIiKi5o+Lqkjsar8NR22IiIgcgeFGYtx2gYiIyLEYbiSWfakMANCaM6WIiIgcguFGYhy5ISIiciyGG4kZp4G3YbghIiJyCIYbCRkMAueLKgBwdWIiIiJHYbiRUK62Ejq9AV5yGaKDfaQuh4iIyC0w3EjI2G/TKsQHCjlXgCYiInIEhhsJGTfMbM01boiIiByG4UZCV2dK8SspIiIiR2G4kRA3zCQiInI8hhsJXR254ddSREREjsJwIyHT6sQcuSEiInIYhhuJaCurcbm8GgC3XiAiInIkhhuJGGdKtfBTwl/lJXE1RERE7oPhRiLnjP02HLUhIiJyKIYbiXCmFBERkXMw3EgkmxtmEhEROQXDjUSMX0txw0wiIiLHYriRyNk/G4rbcOsFIiIih2K4kUCN3oALVyoAsOeGiIjI0RhuJHDxSiX0BgGVlxzhASqpyyEiInIrDDcSyK7TbyOXyySuhoiIyL0w3EjgbFHttgucKUVEROR4DDcSyOZMKSIiIqdhuJFAtmmmFMMNERGRozHcSCCbqxMTERE5DcONiwkhOHJDRETkRAw3LnalvBolVTUAgFYhDDdERESOxnDjYsYNMyMCVVB7KySuhoiIyP0w3LjY1Q0zue0CERGRMzDcuFj2pdo1bjgNnIiIyDmaRLhZvnw5YmNjoVar0bdvXxw4cMDquStXrsSAAQMQEhKCkJAQxMfHN3h+U2MauWEzMRERkVNIHm7Wr1+PKVOmIDk5GYcOHUK3bt2QkJCA/Pz8es/fvXs3Hn/8cXz33XfYv38/NBoN7rnnHly4cMHFlTeOcTdwTgMnIiJyDpkQQkhZQN++fdG7d28sW7YMAGAwGKDRaDBp0iRMnz79utfr9XqEhIRg2bJlGDNmzHXP12q1CAoKQnFxMQIDA2+4fnv1S9mFi8WV2Px8P/RoHeLy1yciImqO7Pn8lnTkRqfT4eDBg4iPjzcdk8vliI+Px/79+216jvLyclRXVyM0NNRZZTpMVY0eOdpKABy5ISIichYvKV+8sLAQer0eERERZscjIiJw8uRJm55j2rRpiI6ONgtIdVVVVaGqqsr0s1arbXzBN+j85QoIAfgpFWjhp5SsDiIiIncmec/NjVi4cCHWrVuHzz//HGq1ut5zUlJSEBQUZHpoNBoXV3lV3Q0zZTKZZHUQERG5M0nDTVhYGBQKBfLy8syO5+XlITIyssFr33zzTSxcuBD/93//h9tuu83qeTNmzEBxcbHpce7cOYfU3hjcdoGIiMj5JA03SqUSPXv2xK5du0zHDAYDdu3ahbi4OKvXvfHGG3j11Vexfft29OrVq8HXUKlUCAwMNHtIhRtmEhEROZ+kPTcAMGXKFCQlJaFXr17o06cPli5dirKyMowbNw4AMGbMGMTExCAlJQUA8K9//Qtz587Fp59+itjYWOTm5gIA/P394e/vL9n7sIVpGngLrk5MRETkLJKHm8TERBQUFGDu3LnIzc1F9+7dsX37dlOTcXZ2NuTyqwNM7733HnQ6HR555BGz50lOTsa8efNcWbrdznHkhoiIyOkkX+fG1aRa50YIgc5zd6CiWo/dU/+K2DCO3hAREdmq2axz40kKSqtQUa2HXAZEB/tIXQ4REZHbYrhxEeNMqehgHyi9eNuJiIichZ+yLsKZUkRERK7BcOMi3DCTiIjINRhuXMQ0U4oL+BERETkVw42LnOXXUkRERC7BcOMixp6bNqGcAk5ERORMDDcuUKHTo6CkdmdyjtwQERE5F8ONCxhHbYJ8vBHk6y1xNURERO6N4cYFOA2ciIjIdRhuXODspTIAnClFRETkCgw3LsANM4mIiFyH4cYFzppmSjHcEBERORvDjQuw54aIiMh1GG6cTG8QOF9UAYA9N0RERK7AcONkedpK6PQGeCtkiArykbocIiIit8dw42TGDTNbhfhCIZdJXA0REZH7Y7hxMuNMKQ37bYiIiFyC4cbJzhbVrnHDmVJERESuwXDjZNnGZmKGGyIiIpdguHGy7D9XJ+bXUkRERK7BcONkxjVu2nAaOBERkUsw3DiRtrIal8urAXDkhoiIyFUYbpwo+89p4GH+SvirvCSuhoiIyDMw3DgRp4ETERG5HsONE3HDTCIiItdjuHEibphJRETkegw3TmTsuWndwk/iSoiIiDwHw40TceSGiIjI9RhunKRab8CFK7WrE3ONGyIiItdhuHGSnCuV0BsEVF5ytPRXSV0OERGRx2C4cRLjhpmtQ30hl8skroaIiMhzMNw4CfttiIiIpMFw4yRXZ0ox3BAREbkSw42TcOSGiIhIGgw3TnL2EncDJyIikgLDjRMIIUz7SnHkhoiIyLUYbpzgcnk1SqpqAACtQhhuiIiIXInhxgmM/TaRgWqovRUSV0NERORZGG6c4Oylq2vcEBERkWsx3DiBqd+GzcREREQux3DjBJwGTkREJB2GGyfgNHAiIiLpMNw4gfFrKQ1HboiIiFyO4cbBqmr0yNFWAgDaMNwQERG5HMONg52/XAEhAD+lAqF+SqnLISIi8jgMNw52dcNMP8hkMomrISIi8jwMNw52daaUj8SVEBEReSaGGwe7OlPKT+JKiIiIPBPDjYNlc6YUERGRpBhuHCy7qHbrBc6UIiIikgbDjQMJIbg6MRERkcQYbhyooKQKldUGyGVATAgbiomIiKTAcONAxlGb6GAfeCt4a4mIiKTAT2AHMs6U4ldSRERE0mG4cSDjyA03zCQiIpIOw40DccNMIiIi6TWJcLN8+XLExsZCrVajb9++OHDgQIPnb9y4EbfccgvUajW6du2Kbdu2uajShp01jtyEcgE/IiIiqUgebtavX48pU6YgOTkZhw4dQrdu3ZCQkID8/Px6z//xxx/x+OOP48knn8Thw4cxbNgwDBs2DMeOHXNx5ZYyC2vXuPFVKiSuhIiIyHPJhBBCygL69u2L3r17Y9myZQAAg8EAjUaDSZMmYfr06RbnJyYmoqysDF999ZXp2F/+8hd0794dK1asuO7rabVaBAUFobi4GIGBgQ57H2v2Z2HOl78DAOQyIGV4VyT2bu2w5yciIvJk9nx+Szpyo9PpcPDgQcTHx5uOyeVyxMfHY//+/fVes3//frPzASAhIcHq+VVVVdBqtWYPR8sprkDylt9NPxsEMHPzMeQUVzj8tYiIiKhhkoabwsJC6PV6REREmB2PiIhAbm5uvdfk5ubadX5KSgqCgoJMD41G45ji68gsLIPhmvEvvRDIKix3+GsRERFRwyTvuXG2GTNmoLi42PQ4d+6cw1+jbZgf5DLzYwqZDLFhnDVFRETkapKGm7CwMCgUCuTl5Zkdz8vLQ2RkZL3XREZG2nW+SqVCYGCg2cPRooJ8kDK8KxSy2oSjkMmwYHgXRAVxCwYiIiJXkzTcKJVK9OzZE7t27TIdMxgM2LVrF+Li4uq9Ji4uzux8APjmm2+snu8qib1bY+/0Qfhs/F+wd/ogNhMTERFJxEvqAqZMmYKkpCT06tULffr0wdKlS1FWVoZx48YBAMaMGYOYmBikpKQAACZPnoyBAwdi8eLFuP/++7Fu3Tr8+uuveP/996V8GwBqR3A4WkNERCQtycNNYmIiCgoKMHfuXOTm5qJ79+7Yvn27qWk4OzsbcvnVAaZ+/frh008/xezZszFz5kzcdNNN+OKLL9ClSxep3gIRERE1IZKvc+NqzlrnhoiIiJyn2axzQ0RERORoDDdERETkVhhuiIiIyK0w3BAREZFbYbghIiIit8JwQ0RERG6F4YaIiIjcCsMNERERuRWGGyIiInIrkm+/4GrGBZm1Wq3ElRAREZGtjJ/btmys4HHhpqSkBACg0WgkroSIiIjsVVJSgqCgoAbP8bi9pQwGAy5evIiAgADIZDKHPrdWq4VGo8G5c+e4b5UT8T67Bu+za/A+uw7vtWs46z4LIVBSUoLo6GizDbXr43EjN3K5HK1atXLqawQGBvJfHBfgfXYN3mfX4H12Hd5r13DGfb7eiI0RG4qJiIjIrTDcEBERkVthuHEglUqF5ORkqFQqqUtxa7zPrsH77Bq8z67De+0aTeE+e1xDMREREbk3jtwQERGRW2G4ISIiIrfCcENERERuheGGiIiI3ArDjZ2WL1+O2NhYqNVq9O3bFwcOHGjw/I0bN+KWW26BWq1G165dsW3bNhdV2rzZc59XrlyJAQMGICQkBCEhIYiPj7/u3xeqZe8/z0br1q2DTCbDsGHDnFugm7D3Pl+5cgUTJkxAVFQUVCoVOnbsyP922MDe+7x06VLcfPPN8PHxgUajwYsvvojKykoXVds8/fDDDxg6dCiio6Mhk8nwxRdfXPea3bt3o0ePHlCpVOjQoQNSU1OdXicE2WzdunVCqVSKVatWid9//12MHz9eBAcHi7y8vHrP37dvn1AoFOKNN94Qx48fF7Nnzxbe3t7i6NGjLq68ebH3Po8cOVIsX75cHD58WJw4cUKMHTtWBAUFifPnz7u48ubF3vtslJmZKWJiYsSAAQPEQw895JpimzF773NVVZXo1auXuO+++8TevXtFZmam2L17t0hLS3Nx5c2Lvff5k08+ESqVSnzyySciMzNT7NixQ0RFRYkXX3zRxZU3L9u2bROzZs0SmzdvFgDE559/3uD5Z86cEb6+vmLKlCni+PHj4p133hEKhUJs377dqXUy3NihT58+YsKECaaf9Xq9iI6OFikpKfWeP2LECHH//febHevbt6945plnnFpnc2fvfb5WTU2NCAgIEB9//LGzSnQLjbnPNTU1ol+/fuKDDz4QSUlJDDc2sPc+v/fee6Jdu3ZCp9O5qkS3YO99njBhgrjrrrvMjk2ZMkX079/fqXW6E1vCzcsvvyxuvfVWs2OJiYkiISHBiZUJwa+lbKTT6XDw4EHEx8ebjsnlcsTHx2P//v31XrN//36z8wEgISHB6vnUuPt8rfLyclRXVyM0NNRZZTZ7jb3Pr7zyCsLDw/Hkk0+6osxmrzH3ecuWLYiLi8OECRMQERGBLl26YMGCBdDr9a4qu9lpzH3u168fDh48aPrq6syZM9i2bRvuu+8+l9TsKaT6HPS4jTMbq7CwEHq9HhEREWbHIyIicPLkyXqvyc3Nrff83Nxcp9XZ3DXmPl9r2rRpiI6OtvgXiq5qzH3eu3cvPvzwQ6SlpbmgQvfQmPt85swZfPvttxg1ahS2bduG9PR0PP/886iurkZycrIrym52GnOfR44cicLCQtxxxx0QQqCmpgbPPvssZs6c6YqSPYa1z0GtVouKigr4+Pg45XU5ckNuZeHChVi3bh0+//xzqNVqqctxGyUlJRg9ejRWrlyJsLAwqctxawaDAeHh4Xj//ffRs2dPJCYmYtasWVixYoXUpbmV3bt3Y8GCBXj33Xdx6NAhbN68GVu3bsWrr74qdWnkABy5sVFYWBgUCgXy8vLMjufl5SEyMrLeayIjI+06nxp3n43efPNNLFy4EDt37sRtt93mzDKbPXvvc0ZGBrKysjB06FDTMYPBAADw8vLCqVOn0L59e+cW3Qw15p/nqKgoeHt7Q6FQmI516tQJubm50Ol0UCqVTq25OWrMfZ4zZw5Gjx6Np556CgDQtWtXlJWV4emnn8asWbMgl/P//R3B2udgYGCg00ZtAI7c2EypVKJnz57YtWuX6ZjBYMCuXbsQFxdX7zVxcXFm5wPAN998Y/V8atx9BoA33ngDr776KrZv345evXq5otRmzd77fMstt+Do0aNIS0szPR588EEMGjQIaWlp0Gg0riy/2WjMP8/9+/dHenq6KTwCwOnTpxEVFcVgY0Vj7nN5eblFgDEGSsEtFx1Gss9Bp7Yru5l169YJlUolUlNTxfHjx8XTTz8tgoODRW5urhBCiNGjR4vp06ebzt+3b5/w8vISb775pjhx4oRITk7mVHAb2HufFy5cKJRKpdi0aZPIyckxPUpKSqR6C82Cvff5WpwtZRt773N2drYICAgQEydOFKdOnRJfffWVCA8PF6+99ppUb6FZsPc+Jycni4CAAPHZZ5+JM2fOiP/7v/8T7du3FyNGjJDqLTQLJSUl4vDhw+Lw4cMCgFiyZIk4fPiwOHv2rBBCiOnTp4vRo0ebzjdOBf/nP/8pTpw4IZYvX86p4E3RO++8I1q3bi2USqXo06eP+Omnn0y/GzhwoEhKSjI7f8OGDaJjx45CqVSKW2+9VWzdutXFFTdP9tznNm3aCAAWj+TkZNcX3szY+89zXQw3trP3Pv/444+ib9++QqVSiXbt2onXX39d1NTUuLjq5see+1xdXS3mzZsn2rdvL9RqtdBoNOL5558Xly9fdn3hzch3331X739vjfc2KSlJDBw40OKa7t27C6VSKdq1ayc++ugjp9cpE4Ljb0REROQ+2HNDREREboXhhoiIiNwKww0RERG5FYYbIiIicisMN0RERORWGG6IiIjIrTDcEBERkVthuCEiM6mpqQgODpa6jEaTyWT44osvGjxn7NixGDZsmEvqISLXY7ghckNjx46FTCazeKSnp0tdGlJTU031yOVytGrVCuPGjUN+fr5Dnj8nJwf33nsvACArKwsymQxpaWlm57z99ttITU11yOtZM2/ePNP7VCgU0Gg0ePrpp1FUVGTX8zCIEdmPu4ITuakhQ4bgo48+MjvWsmVLiaoxFxgYiFOnTsFgMOC3337DuHHjcPHiRezYseOGn/t6u8cDQFBQ0A2/ji1uvfVW7Ny5E3q9HidOnMATTzyB4uJirF+/3iWvT+SpOHJD5KZUKhUiIyPNHgqFAkuWLEHXrl3h5+cHjUaD559/HqWlpVaf57fffsOgQYMQEBCAwMBA9OzZE7/++qvp93v37sWAAQPg4+MDjUaDF154AWVlZQ3WJpPJEBkZiejoaNx777144YUXsHPnTlRUVMBgMOCVV15Bq1atoFKp0L17d2zfvt10rU6nw8SJExEVFQW1Wo02bdogJSXF7LmNX0u1bdsWAHD77bdDJpPhr3/9KwDz0ZD3338f0dHRZrtwA8BDDz2EJ554wvTzl19+iR49ekCtVqNdu3aYP38+ampqGnyfXl5eiIyMRExMDOLj4/Hoo4/im2++Mf1er9fjySefRNu2beHj44Obb74Zb7/9tun38+bNw8cff4wvv/zSNAq0e/duAMC5c+cwYsQIBAcHIzQ0FA899BCysrIarIfIUzDcEHkYuVyOf//73/j999/x8ccf49tvv8XLL79s9fxRo0ahVatW+OWXX3Dw4EFMnz4d3t7eAICMjAwMGTIEDz/8MI4cOYL169dj7969mDhxol01+fj4wGAwoKamBm+//TYWL16MN998E0eOHEFCQgIefPBB/PHHHwCAf//739iyZQs2bNiAU6dO4ZNPPkFsbGy9z3vgwAEAwM6dO5GTk4PNmzdbnPPoo4/i0qVL+O6770zHioqKsH37dowaNQoAsGfPHowZMwaTJ0/G8ePH8Z///Aepqal4/fXXbX6PWVlZ2LFjB5RKpemYwWBAq1atsHHjRhw/fhxz587FzJkzsWHDBgDA1KlTMWLECAwZMgQ5OTnIyclBv379UF1djYSEBAQEBGDPnj3Yt28f/P39MWTIEOh0OptrInJbTt+ak4hcLikpSSgUCuHn52d6PPLII/Weu3HjRtGiRQvTzx999JEICgoy/RwQECBSU1PrvfbJJ58UTz/9tNmxPXv2CLlcLioqKuq95trnP336tOjYsaPo1auXEEKI6Oho8frrr5td07t3b/H8888LIYSYNGmSuOuuu4TBYKj3+QGIzz//XAghRGZmpgAgDh8+bHbOtTuaP/TQQ+KJJ54w/fyf//xHREdHC71eL4QQYvDgwWLBggVmz7FmzRoRFRVVbw1CCJGcnCzkcrnw8/MTarXatHvykiVLrF4jhBATJkwQDz/8sNVaja998803m92Dqqoq4ePjI3bs2NHg8xN5AvbcELmpQYMG4b333jP97OfnB6B2FCMlJQUnT56EVqtFTU0NKisrUV5eDl9fX4vnmTJlCp566imsWbPG9NVK+/btAdR+ZXXkyBF88sknpvOFEDAYDMjMzESnTp3qra24uBj+/v4wGAyorKzEHXfcgQ8++ABarRYXL15E//79zc7v378/fvvtNwC1XyndfffduPnmmzFkyBA88MADuOeee27oXo0aNQrjx4/Hu+++C5VKhU8++QSPPfYY5HK56X3u27fPbKRGr9c3eN8A4Oabb8aWLVtQWVmJtWvXIi0tDZMmTTI7Z/ny5Vi1ahWys7NRUVEBnU6H7t27N1jvb7/9hvT0dAQEBJgdr6ysREZGRiPuAJF7YbghclN+fn7o0KGD2bGsrCw88MADeO655/D6668jNDQUe/fuxZNPPgmdTlfvh/S8efMwcuRIbN26FV9//TWSk5Oxbt06/O1vf0NpaSmeeeYZvPDCCxbXtW7d2mptAQEBOHToEORyOaKiouDj4wMA0Gq1131fPXr0QGZmJr7++mvs3LkTI0aMQHx8PDZt2nTda60ZOnQohBDYunUrevfujT179uCtt94y/b60tBTz58/H8OHDLa5Vq9VWn1epVJr+HixcuBD3338/5s+fj1dffRUAsG7dOkydOhWLFy9GXFwcAgICsGjRIvz8888N1ltaWoqePXuahUqjptI0TiQlhhsiD3Lw4EEYDAYsXrzYNCph7O9oSMeOHdGxY0e8+OKLePzxx/HRRx/hb3/7G3r06IHjx49bhKjrkcvl9V4TGBiI6Oho7Nu3DwMHDjQd37dvH/r06WN2XmJiIhITE/HII49gyJAhKCoqQmhoqNnzGftb9Hp9g/Wo1WoMHz4cn3zyCdLT03HzzTejR48ept/36NEDp06dsvt9Xmv27Nm466678Nxzz5neZ79+/fD888+bzrl25EWpVFrU36NHD6xfvx7h4eEIDAy8oZqI3BEbiok8SIcOHVBdXY133nkHZ86cwZo1a7BixQqr51dUVGDixInYvXs3zp49i3379uGXX34xfd00bdo0/Pjjj5g4cSLS0tLwxx9/4Msvv7S7obiuf/7zn/jXv/6F9evX49SpU5g+fTrS0tIwefJkAMCSJUvw2Wef4eTJkzh9+jQ2btyIyMjIehceDA8Ph4+PD7Zv3468vDwUFxdbfd1Ro0Zh69atWLVqlamR2Gju3LlYvXo15s+fj99//x0nTpzAunXrMHv2bLveW1xcHG677TYsWLAAAHDTTTfh119/xY4dO3D69GnMmTMHv/zyi9k1sbGxOHLkCE6dOoXCwkJUV1dj1KhRCAsLw0MPPYQ9e/YgMzMTu3fvxgsvvIDz58/bVRORW5K66YeIHK++JlSjJUuWiKioKOHj4yMSEhLE6tWrBQBx+fJlIYR5w29VVZV47LHHhEajEUqlUkRHR4uJEyeaNQsfOHBA3H333cLf31/4+fmJ2267zaIhuK5rG4qvpdfrxbx580RMTIzw9vYW3bp1E19//bXp9++//77o3r278PPzE4GBgWLw4MHi0KFDpt+jTkOxEEKsXLlSaDQaIZfLxcCBA63eH71eL6KiogQAkZGRYVHX9u3bRb9+/YSPj48IDAwUffr0Ee+//77V95GcnCy6detmcfyzzz4TKpVKZGdni8rKSjF27FgRFBQkgoODxXPPPSemT59udl1+fr7p/gIQ3333nRBCiJycHDFmzBgRFhYmVCqVaNeunRg/frwoLi62WhORp5AJIYS08YqIiIjIcfi1FBEREbkVhhsiIiJyKww3RERE5FYYboiIiMitMNwQERGRW2G4ISIiIrfCcENERERuheGGiIiI3ArDDREREbkVhhsiIiJyKww3RERE5FYYboiIiMit/D/7uFD8/MQBvwAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "\n",
    "fpr, tpr, thresholds = roc_curve(targets_test, test_predictions)\n",
    "\n",
    "plt.plot(fpr, tpr, marker='.')\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC Curve')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import precision_recall_curve\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "precision, recall, thresholds = precision_recall_curve(targets_test, test_predictions)\n",
    "\n",
    "plt.plot(recall, precision, marker='.')\n",
    "plt.xlabel('Recall')\n",
    "plt.ylabel('Precision')\n",
    "plt.title('Precision-Recall Curve')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## ML Flow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "id": "VR4vWvUkQa48"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024/12/17 18:04:09 WARNING mlflow.keras.save: You are saving a Keras model without specifying model signature.\n",
      "2024/12/17 18:04:21 WARNING mlflow.models.model: Model logged without a signature and input example. Please set `input_example` parameter when logging the model to auto infer the model signature.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Modelo salvo com sucesso no MLflow!\n",
      "🏃 View run calm-bee-638 at: http://127.0.0.1:5000/#/experiments/270529103012719304/runs/4232eac3c1be4c369789b58d3d4f0112\n",
      "🧪 View experiment at: http://127.0.0.1:5000/#/experiments/270529103012719304\n"
     ]
    }
   ],
   "source": [
    "import mlflow\n",
    "import mlflow.keras  # Para modelos Keras\n",
    "\n",
    "\n",
    "###################################### REGISTRO NO MLFLOW ################################################\n",
    "\n",
    "# Set our tracking server URI for logging\n",
    "mlflow.set_tracking_uri('http://127.0.0.1:5000')\n",
    "\n",
    "# Configurar o MLflow para registrar a execução do RandomForestClassifier\n",
    "mlflow.set_experiment('fraud_detection_experiment') # Define o nome do experimento\n",
    "\n",
    "with mlflow.start_run():\n",
    "\n",
    "    # Logando os parâmetros\n",
    "    mlflow.log_param('epochs', 10)\n",
    "    mlflow.log_param('batch_size', 32)\n",
    "    mlflow.log_param('optimizer', 'adam')\n",
    "    mlflow.log_param('loss_function', 'binary_crossentropy')\n",
    "    \n",
    "    # Logando a métrica (acurácia)\n",
    "    #mlflow.log_metric('accuracy', accuracy[1])\n",
    "\n",
    "    # Logando a arquitetura do modelo\n",
    "    #model_json = model.to_json()\n",
    "    #mlflow.log_artifact(model_json, 'model_architecture.json')\n",
    "\n",
    "    # Logando o modelo (para salvar o modelo inteiro)\n",
    "    mlflow.keras.log_model(model, 'fraud_detection_model')\n",
    "\n",
    "    print(\"Modelo salvo com sucesso no MLflow!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "authorship_tag": "ABX9TyMOauGt7LFGUAh1c82no8+H",
   "mount_file_id": "114AIwxZ6aKzzVfdHWqEjmWH9YKK6_lJE",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "neural_env",
   "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": "3.9.21"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}