basil_fraud_or_not.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "\u001b[1m7108/7108\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 322us/step - loss: 0.6082 - val_loss: 0.4525\n",
      "Epoch 2/50\n",
      "\u001b[1m7108/7108\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 314us/step - loss: 0.4466 - val_loss: 0.4545\n",
      "Epoch 3/50\n",
      "\u001b[1m7108/7108\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 309us/step - loss: 0.4483 - val_loss: 0.4534\n",
      "Epoch 4/50\n",
      "\u001b[1m7108/7108\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 314us/step - loss: 0.4333 - val_loss: 0.4523\n",
      "Epoch 5/50\n",
      "\u001b[1m7108/7108\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 312us/step - loss: 0.4368 - val_loss: 0.4520\n",
      "Epoch 6/50\n",
      "\u001b[1m7108/7108\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 313us/step - loss: 0.4520 - val_loss: 0.4529\n",
      "Epoch 7/50\n",
      "\u001b[1m7108/7108\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 313us/step - loss: 0.4436 - val_loss: 0.4545\n",
      "Epoch 8/50\n",
      "\u001b[1m7108/7108\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 317us/step - loss: 0.4398 - val_loss: 0.4543\n",
      "Epoch 9/50\n",
      "\u001b[1m7108/7108\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 313us/step - loss: 0.4385 - val_loss: 0.4540\n",
      "Epoch 10/50\n",
      "\u001b[1m7108/7108\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 315us/step - loss: 0.4414 - val_loss: 0.4544\n",
      "\u001b[1m8901/8901\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 219us/step\n",
      "Accuracy: 0.9513179100232789\n",
      "Precision: 0.03047538796432835\n",
      "Recall: 0.8821138211382114\n",
      "F1-Score: 0.05891536007601982\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.layers import Input, Dense\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "\n",
    "# Load the dataset\n",
    "data = pd.read_csv(\"creditcard.csv\") \n",
    "\n",
    "# Separate features and labels\n",
    "X = data.drop(['Class'], axis=1)\n",
    "y = data['Class']\n",
    "\n",
    "# Standardize the features\n",
    "scaler = StandardScaler()\n",
    "X_scaled = scaler.fit_transform(X)\n",
    "\n",
    "# Split into training and testing sets (using only non-fraudulent data for training)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_scaled[y == 0], y[y == 0], test_size=0.2, random_state=42)\n",
    "\n",
    "# Define the autoencoder architecture\n",
    "input_dim = X_train.shape[1]\n",
    "encoding_dim = 14  # Adjust this as needed\n",
    "\n",
    "input_layer = Input(shape=(input_dim,))\n",
    "encoder = Dense(encoding_dim, activation='relu')(input_layer)\n",
    "decoder = Dense(input_dim, activation='linear')(encoder)\n",
    "\n",
    "autoencoder = Model(inputs=input_layer, outputs=decoder)\n",
    "autoencoder.compile(optimizer='adam', loss='mse')\n",
    "\n",
    "# Train the autoencoder\n",
    "early_stopping = EarlyStopping(monitor='val_loss', patience=5, restore_best_weights=True)\n",
    "autoencoder.fit(X_train, X_train, epochs=50, batch_size=32, validation_data=(X_test, X_test), callbacks=[early_stopping])\n",
    "\n",
    "# Calculate reconstruction errors for the entire dataset\n",
    "reconstructions = autoencoder.predict(X_scaled)\n",
    "mse = np.mean(np.power(X_scaled - reconstructions, 2), axis=1)\n",
    "\n",
    "# Set a threshold for anomaly detection (you'll need to tune this)\n",
    "threshold = np.percentile(mse, 95)  # Adjust this percentile as needed\n",
    "\n",
    "# Identify anomalies\n",
    "anomalies = (mse > threshold)\n",
    "\n",
    "# Evaluate the model (using the original labels)\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score\n",
    "\n",
    "print(\"Accuracy:\", accuracy_score(y, anomalies))\n",
    "print(\"Precision:\", precision_score(y, anomalies))\n",
    "print(\"Recall:\", recall_score(y, anomalies))\n",
    "print(\"F1-Score:\", f1_score(y, anomalies))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import precision_recall_curve, auc\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "\n",
    "# Calculate precision, recall, and thresholds\n",
    "precision, recall, thresholds = precision_recall_curve(y, mse)\n",
    "\n",
    "# Calculate AUPRC\n",
    "auprc = auc(recall, precision)\n",
    "\n",
    "# Create the plot\n",
    "plt.figure(figsize=(8, 6))\n",
    "plt.plot(recall, precision, label=f'AUPRC = {auprc:.2f}')\n",
    "plt.xlabel('Recall')\n",
    "plt.ylabel('Precision')\n",
    "plt.title('Precision-Recall Curve for Anomaly Detection')\n",
    "plt.legend(loc='lower left')\n",
    "plt.text(0.1, 0.9, f'AUPRC: {auprc:.2f}', fontsize=12)\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}