03_AutoXGB.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "9hM6lgcUnkAR"
   },
   "source": [
    "# Credit Card Fraud Detection with AutoXGB\n",
    "\n",
    "## Part 3 - Running AutoXGB\n",
    "___\n",
    "Reference: https://github.com/abhishekkrthakur/autoxgb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "dXHyUcjWnkAW"
   },
   "source": [
    "### Install AutoXGB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "8IAuECIqr1ux",
    "outputId": "8599cadc-b425-4ca4-a8bc-1214b6ff2f74"
   },
   "outputs": [],
   "source": [
    "# !pip install autoxgb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "R5l1FTl8nkAY"
   },
   "source": [
    "### Import dependencies and data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "kFMwklZbnkAY"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from autoxgb import AutoXGB\n",
    "import sklearn.metrics as metrics\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "from IPython.display import display\n",
    "pd.set_option('display.max_columns', None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "KBvwrUgOr1uz"
   },
   "source": [
    "### Run AutoXGB Training and Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "zXAZLNAonkAY",
    "outputId": "7cc1fb70-8afe-47b8-b41c-fb12752419fd"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2021-11-28 09:41:18.180 | INFO     | autoxgb.autoxgb:__post_init__:42 - Output directory: output\n",
      "2021-11-28 09:41:18.186 | INFO     | autoxgb.autoxgb:_process_data:149 - Reading training data\n",
      "2021-11-28 09:41:18.389 | INFO     | autoxgb.utils:reduce_memory_usage:50 - Mem. usage decreased to 2.28 Mb (77.3% reduction)\n",
      "2021-11-28 09:41:18.397 | INFO     | autoxgb.autoxgb:_determine_problem_type:140 - Problem type: binary_classification\n",
      "2021-11-28 09:41:18.530 | INFO     | autoxgb.utils:reduce_memory_usage:50 - Mem. usage decreased to 1.86 Mb (77.3% reduction)\n",
      "2021-11-28 09:41:18.533 | INFO     | autoxgb.autoxgb:_create_folds:58 - Creating folds\n",
      "2021-11-28 09:41:18.562 | INFO     | autoxgb.autoxgb:_process_data:170 - Encoding target(s)\n",
      "2021-11-28 09:41:18.571 | INFO     | autoxgb.autoxgb:_process_data:195 - Found 0 categorical features.\n",
      "2021-11-28 09:41:18.821 | INFO     | autoxgb.autoxgb:_process_data:236 - Model config: train_filename='data/processed/train.csv' test_filename='data/processed/test.csv' idx='id' targets=['TX_FRAUD'] problem_type=<ProblemType.binary_classification: 1> output='output' features=['TX_AMOUNT', 'TX_DURING_WEEKEND', 'TX_DURING_NIGHT', 'CUSTOMER_ID_NB_TX_1DAY_WINDOW', 'CUSTOMER_ID_NB_TX_7DAY_WINDOW', 'CUSTOMER_ID_NB_TX_30DAY_WINDOW', 'CUSTOMER_ID_AVG_AMOUNT_1DAY_WINDOW', 'CUSTOMER_ID_AVG_AMOUNT_7DAY_WINDOW', 'CUSTOMER_ID_AVG_AMOUNT_30DAY_WINDOW', 'TERMINAL_ID_NB_TX_1DAY_WINDOW', 'TERMINAL_ID_NB_TX_7DAY_WINDOW', 'TERMINAL_ID_NB_TX_30DAY_WINDOW', 'TERMINAL_ID_RISK_1DAY_WINDOW', 'TERMINAL_ID_RISK_7DAY_WINDOW', 'TERMINAL_ID_RISK_30DAY_WINDOW'] num_folds=5 use_gpu=True seed=42 categorical_features=[] num_trials=200 time_limit=400 fast=False\n",
      "2021-11-28 09:41:18.823 | INFO     | autoxgb.autoxgb:_process_data:237 - Saving model config\n",
      "2021-11-28 09:41:18.830 | INFO     | autoxgb.autoxgb:_process_data:241 - Saving encoders\n",
      "\u001b[32m[I 2021-11-28 09:41:19,144]\u001b[0m A new study created in RDB with name: autoxgb\u001b[0m\n",
      "2021-11-28 09:41:45.520 | INFO     | autoxgb.utils:optimize:192 - Metrics: {'auc': 0.98490056879468, 'logloss': 0.07456094019196087, 'f1': 0.9721964827028359, 'accuracy': 0.981850102918029, 'precision': 0.9932538361174178, 'recall': 0.9520159825644752}\n",
      "\u001b[32m[I 2021-11-28 09:41:45,572]\u001b[0m Trial 0 finished with value: 0.07456094019196087 and parameters: {'learning_rate': 0.24006160068298701, 'reg_lambda': 13.630157010253162, 'reg_alpha': 1.574522470129688e-05, 'subsample': 0.8259568497932395, 'colsample_bytree': 0.807743847546043, 'max_depth': 2, 'early_stopping_rounds': 270, 'n_estimators': 7000}. Best is trial 0 with value: 0.07456094019196087.\u001b[0m\n",
      "2021-11-28 09:43:01.950 | INFO     | autoxgb.utils:optimize:192 - Metrics: {'auc': 0.9747686884261517, 'logloss': 0.14327487733495906, 'f1': 0.9279715564396345, 'accuracy': 0.9544012592323525, 'precision': 0.9799660786846698, 'recall': 0.8812204867417363}\n",
      "\u001b[32m[I 2021-11-28 09:43:01,992]\u001b[0m Trial 1 finished with value: 0.14327487733495906 and parameters: {'learning_rate': 0.01245522288702717, 'reg_lambda': 5.887739136109787e-08, 'reg_alpha': 0.24788571477310173, 'subsample': 0.563117602108261, 'colsample_bytree': 0.9096620618063324, 'max_depth': 1, 'early_stopping_rounds': 259, 'n_estimators': 7000}. Best is trial 0 with value: 0.07456094019196087.\u001b[0m\n",
      "2021-11-28 09:45:45.600 | INFO     | autoxgb.utils:optimize:192 - Metrics: {'auc': 0.9791629403659501, 'logloss': 0.1184076415235095, 'f1': 0.9461597425094015, 'accuracy': 0.9655527303547645, 'precision': 0.9875224927048091, 'recall': 0.908136578278242}\n",
      "\u001b[32m[I 2021-11-28 09:45:45,645]\u001b[0m Trial 2 finished with value: 0.1184076415235095 and parameters: {'learning_rate': 0.01572678361680493, 'reg_lambda': 6.524292517102156e-05, 'reg_alpha': 9.762431593599686e-08, 'subsample': 0.9455185813483441, 'colsample_bytree': 0.36353458421907936, 'max_depth': 1, 'early_stopping_rounds': 419, 'n_estimators': 15000}. Best is trial 0 with value: 0.07456094019196087.\u001b[0m\n",
      "2021-11-28 09:48:13.179 | INFO     | autoxgb.utils:optimize:192 - Metrics: {'auc': 0.9865799209368931, 'logloss': 0.07096395374182347, 'f1': 0.9734676585227223, 'accuracy': 0.982661339145175, 'precision': 0.9934588885110838, 'recall': 0.9542680711950599}\n",
      "\u001b[32m[I 2021-11-28 09:48:13,224]\u001b[0m Trial 3 finished with value: 0.07096395374182347 and parameters: {'learning_rate': 0.010054636551750887, 'reg_lambda': 0.000637733927666517, 'reg_alpha': 1.5793759643443387e-07, 'subsample': 0.7189011968067806, 'colsample_bytree': 0.7002312718865316, 'max_depth': 6, 'early_stopping_rounds': 251, 'n_estimators': 7000}. Best is trial 3 with value: 0.07096395374182347.\u001b[0m\n",
      "2021-11-28 09:48:13.237 | INFO     | autoxgb.autoxgb:train:248 - Training complete\n",
      "2021-11-28 09:48:13.238 | INFO     | autoxgb.autoxgb:predict:252 - Creating OOF and test predictions\n",
      "2021-11-28 09:48:13.243 | INFO     | autoxgb.utils:predict_model:238 - Training and predicting for fold 0\n",
      "2021-11-28 09:48:46.370 | INFO     | autoxgb.utils:predict_model:333 - Fold 0 done!\n",
      "2021-11-28 09:48:46.372 | INFO     | autoxgb.utils:predict_model:238 - Training and predicting for fold 1\n",
      "2021-11-28 09:49:16.857 | INFO     | autoxgb.utils:predict_model:333 - Fold 1 done!\n",
      "2021-11-28 09:49:16.859 | INFO     | autoxgb.utils:predict_model:238 - Training and predicting for fold 2\n",
      "2021-11-28 09:49:47.815 | INFO     | autoxgb.utils:predict_model:333 - Fold 2 done!\n",
      "2021-11-28 09:49:47.817 | INFO     | autoxgb.utils:predict_model:238 - Training and predicting for fold 3\n",
      "2021-11-28 09:50:18.459 | INFO     | autoxgb.utils:predict_model:333 - Fold 3 done!\n",
      "2021-11-28 09:50:18.461 | INFO     | autoxgb.utils:predict_model:238 - Training and predicting for fold 4\n",
      "2021-11-28 09:50:44.901 | INFO     | autoxgb.utils:predict_model:333 - Fold 4 done!\n",
      "2021-11-28 09:50:44.903 | INFO     | autoxgb.utils:predict_model:336 - Metrics: {'auc': 0.9865799209368931, 'logloss': 0.07096395374182347, 'f1': 0.9734676585227223, 'accuracy': 0.982661339145175, 'precision': 0.9934588885110838, 'recall': 0.9542680711950599}\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 9min 25s, sys: 4.23 s, total: 9min 29s\n",
      "Wall time: 9min 27s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "# Define required parameters\n",
    "train_filename = \"data/processed/train.csv\" # Path to training dataset\n",
    "output = \"output_autoxgb\" # Name of output folder\n",
    "\n",
    "# Set optional parameters\n",
    "test_filename = \"data/processed/test.csv\"\n",
    "task = \"classification\"\n",
    "targets = [\"TX_FRAUD\"]\n",
    "use_gpu = True\n",
    "num_folds = 5 # Number of CV folds\n",
    "seed = 42\n",
    "num_trials = 200\n",
    "time_limit = 400\n",
    "fast = False\n",
    "\n",
    "# Start AutoXGB training\n",
    "axgb = AutoXGB(\n",
    "            train_filename=train_filename,\n",
    "            output=output,\n",
    "            test_filename=test_filename,\n",
    "            task=task,\n",
    "            targets=targets,\n",
    "            use_gpu=use_gpu,\n",
    "            num_folds=num_folds,\n",
    "            seed=seed,\n",
    "            num_trials=num_trials,\n",
    "            time_limit=time_limit,\n",
    "            fast=fast,\n",
    ")\n",
    "\n",
    "axgb.train()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TXXpTZiCr1u2"
   },
   "source": [
    "### Setup test dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "id": "tmmFJvpNnkAa"
   },
   "outputs": [],
   "source": [
    "# Import dataset and define target variable\n",
    "test_df = pd.read_csv('data/processed/test.csv')\n",
    "target = 'TX_FRAUD'\n",
    "\n",
    "# Get y true labels\n",
    "y_test = test_df[target]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ruUx-Jdor1u3"
   },
   "source": [
    "### Get predictions on test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "id": "3VlZlfcrnkAb",
    "outputId": "ea5c41a8-83c0-4ce9-eb20-3db36c8ee83c"
   },
   "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>id</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.952505</td>\n",
       "      <td>0.047495</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.984080</td>\n",
       "      <td>0.015920</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0.996838</td>\n",
       "      <td>0.003162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0.981726</td>\n",
       "      <td>0.018274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0.000083</td>\n",
       "      <td>0.999917</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   id         0         1\n",
       "0   0  0.952505  0.047495\n",
       "1   1  0.984080  0.015920\n",
       "2   2  0.996838  0.003162\n",
       "3   3  0.981726  0.018274\n",
       "4   4  0.000083  0.999917"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# View probability predictions from AutoXGB\n",
    "preds_proba = pd.read_csv('output/test_predictions.csv')\n",
    "preds_proba.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "vvDSa8e6wUXQ",
    "outputId": "5cf20c08-b19d-48de-b9f9-95a1ee05dfa4"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0        0.047495\n",
       "1        0.015920\n",
       "2        0.003162\n",
       "3        0.018274\n",
       "4        0.999917\n",
       "           ...   \n",
       "67381    0.996280\n",
       "67382    0.027509\n",
       "67383    0.017280\n",
       "67384    0.011525\n",
       "67385    0.028157\n",
       "Name: 1, Length: 67386, dtype: float64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get prediction probabilities for label 1\n",
    "y_preds_proba = preds_proba.iloc[:,2]\n",
    "y_preds_proba"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 206
    },
    "id": "E4IzCOhe4DTu",
    "outputId": "9bf0b017-6c04-41f1-94fa-879ef2d0ac05"
   },
   "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>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   label\n",
       "0      0\n",
       "1      0\n",
       "2      0\n",
       "3      0\n",
       "4      1"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Set threshold (default = 0.5)\n",
    "threshold = 0.5\n",
    "\n",
    "# Generate predicted labels\n",
    "preds = preds_proba.copy()\n",
    "preds['label'] = preds['1'].apply(lambda x: 0 if x < threshold else 1)\n",
    "preds = preds.iloc[: , 3:] # Drop unnecessary columns\n",
    "preds.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "lGLNllOvr1u5"
   },
   "source": [
    "### Review performance metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "id": "m7jWIkC9r1u5"
   },
   "outputs": [],
   "source": [
    "# Define function to consolidate and display performance metrics\n",
    "def gen_performance_metrics(y_true, y_preds, y_preds_proba, model_name, round_to=3):\n",
    "    accuracy = round(metrics.accuracy_score(y_true, y_preds), round_to)\n",
    "    f1_score = round(metrics.f1_score(y_true, y_preds), round_to)\n",
    "    roc_auc = round(metrics.roc_auc_score(y_true, y_preds_proba), round_to)\n",
    "    avg_precision = round(metrics.average_precision_score(y_true, y_preds_proba), round_to)\n",
    "\n",
    "    df = pd.DataFrame([[model_name, accuracy, f1_score, roc_auc, avg_precision]], \n",
    "                        columns=['Model', 'Accuracy', 'F1 Score', 'ROC AUC Score', 'Average Precision'])\n",
    "    \n",
    "    display(df)\n",
    "    display(metrics.ConfusionMatrixDisplay.from_predictions(y_true, y_preds))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 360
    },
    "id": "uKcL610Zr1u5",
    "outputId": "a1ba1efd-d81d-43fc-d8a2-bbd0eab767dd"
   },
   "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>Model</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>F1 Score</th>\n",
       "      <th>ROC AUC Score</th>\n",
       "      <th>Average Precision</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>AutoXGB</td>\n",
       "      <td>0.995</td>\n",
       "      <td>0.723</td>\n",
       "      <td>0.914</td>\n",
       "      <td>0.782</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Model  Accuracy  F1 Score  ROC AUC Score  Average Precision\n",
       "0  AutoXGB     0.995     0.723          0.914              0.782"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x7f84ed9c0f10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Show performance assessment\n",
    "gen_performance_metrics(y_test, preds, y_preds_proba, 'AutoXGB')"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "03_AutoXGB.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}