3_Neural_Network.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "deda8a21",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importing required libraries\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from matplotlib import pyplot as plt \n",
    "import seaborn as sns\n",
    "from sklearn.metrics import f1_score , accuracy_score ,precision_score , recall_score , confusion_matrix , precision_recall_curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "77c66585",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importing required tensorflow functions\n",
    "import tensorflow as tf\n",
    "tf.random.set_seed(221)                             # pip install tensorflow\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense\n",
    "from tensorflow.keras.losses import BinaryCrossentropy\n",
    "from tensorflow.keras.optimizers import Adam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "369e4536",
   "metadata": {},
   "outputs": [],
   "source": [
    "def prediction_maker(Y_hat,Y_test):\n",
    "    precision, recall, thresholds = precision_recall_curve(Y_test, Y_hat)\n",
    "    fscore = (2 * precision * recall) / (precision + recall)\n",
    "    ix = np.argmax(fscore) \n",
    "\n",
    "    Y_pred_new = []\n",
    "    for i in range(len(Y_hat)) : \n",
    "        if Y_hat[i] < thresholds[ix] :\n",
    "            Y_pred_new.append(0)\n",
    "        else :  \n",
    "            Y_pred_new.append(1) \n",
    "\n",
    "    return Y_pred_new , thresholds[ix]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "a21de6b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "def confusion_matrix_plot(Y_test,y_pred):\n",
    "    plt.figure(figsize=(6,6))\n",
    "    data = confusion_matrix(Y_test,y_pred) \n",
    "    l = ['Normal','Fraud']\n",
    "    sns.heatmap(data,annot= True,annot_kws= {'size':18},fmt= 'd',linewidths= .5,cmap= 'vlag',square= True,xticklabels= l ,yticklabels= l)\n",
    "    plt.title('Confusion Matrix')\n",
    "    plt.ylabel('Actual transaction')\n",
    "    plt.xlabel('Predicted transaction')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "29232e7f",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = pd.read_csv(\"C:/Users/prane/Downloads/Github/CREDIT-CARD-FRAUD-DETECTION/X_train.csv\").to_numpy()\n",
    "Y_train = pd.read_csv(\"C:/Users/prane/Downloads/Github/CREDIT-CARD-FRAUD-DETECTION/Y_train.csv\").to_numpy().ravel()\n",
    "X_test = pd.read_csv(\"C:/Users/prane/Downloads/Github/CREDIT-CARD-FRAUD-DETECTION/X_test.csv\").to_numpy()  \n",
    "Y_test = pd.read_csv(\"C:/Users/prane/Downloads/Github/CREDIT-CARD-FRAUD-DETECTION/Y_test.csv\").to_numpy().ravel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "d245fef1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[(220530, 29), (220530,)], [(55133, 29), (55133,)]]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[[ X_train.shape , Y_train.shape ],[ X_test.shape , Y_test.shape ]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "a371d022",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating the layers of Neural Network.\n",
    "model = Sequential([\n",
    "                        tf.keras.Input(shape=(29,)),\n",
    "                        Dense(units=20,activation='relu'),\n",
    "                        Dense(units=7,activation='relu'),\n",
    "                        Dense(units=5,activation='relu'),\n",
    "                        Dense(units=1,activation='sigmoid'),\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "43944083",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Setting the Cost Function and Optimisation method.\n",
    "model.compile(\n",
    "    loss=tf.keras.losses.BinaryCrossentropy(),\n",
    "    optimizer=tf.keras.optimizers.Adam(0.001),\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "4fcd7487",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/24\n",
      "6892/6892 [==============================] - 11s 1ms/step - loss: 0.0088\n",
      "Epoch 2/24\n",
      "6892/6892 [==============================] - 10s 1ms/step - loss: 0.0033\n",
      "Epoch 3/24\n",
      "6892/6892 [==============================] - 13s 2ms/step - loss: 0.0030\n",
      "Epoch 4/24\n",
      "6892/6892 [==============================] - 17s 2ms/step - loss: 0.0028\n",
      "Epoch 5/24\n",
      "6892/6892 [==============================] - 14s 2ms/step - loss: 0.0026\n",
      "Epoch 6/24\n",
      "6892/6892 [==============================] - 13s 2ms/step - loss: 0.0025\n",
      "Epoch 7/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0024\n",
      "Epoch 8/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0024\n",
      "Epoch 9/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0022\n",
      "Epoch 10/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0022\n",
      "Epoch 11/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0022\n",
      "Epoch 12/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0021\n",
      "Epoch 13/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0020\n",
      "Epoch 14/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0020\n",
      "Epoch 15/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0020\n",
      "Epoch 16/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0019\n",
      "Epoch 17/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0019\n",
      "Epoch 18/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0019\n",
      "Epoch 19/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0019\n",
      "Epoch 20/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0018\n",
      "Epoch 21/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0018\n",
      "Epoch 22/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0018\n",
      "Epoch 23/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0016\n",
      "Epoch 24/24\n",
      "6892/6892 [==============================] - 12s 2ms/step - loss: 0.0017\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x144ecea9e40>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Fitting the train data to neural network model\n",
    "model.fit(X_train,Y_train,epochs=24)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "1e55c93d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1723/1723 [==============================] - 3s 2ms/step\n"
     ]
    }
   ],
   "source": [
    "# Getting the output values for the test data\n",
    "y_ = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "b99081a7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Converting the output data to 1 and 0 based on threshold of 0.5\n",
    "y = tf.round(y_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "2766baac",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Changing the threshold value of 0.5 to another suited value to obtain better results\n",
    "Y , threshold= prediction_maker(y_,Y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "ab42a2dd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.40638655"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# The value of threshold\n",
    "threshold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "73c1e4a3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9994558612809026"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(Y_test , y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "4c8d0de0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.999492137195509"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(Y_test , Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "d31c5d35",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9450056973961158"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision_score(Y_test , y , average = 'macro')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "f0b1d369",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9194130635151123"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision_score(Y_test , Y , average = 'macro')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "e5068440",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8734905518936931"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(Y_test , y , average = 'macro')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "c53d1906",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9194130635151123"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(Y_test , Y , average = 'macro')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "51752367",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9061137676420904"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f1_score(Y_test , y , average = 'macro')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "03dd1d03",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9194130635151123"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f1_score(Y_test , Y , average = 'macro')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "05b1b4f9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusion_matrix_plot(Y_test,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "4d733265",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusion_matrix_plot(Y_test,Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "06822b83",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Saving the weights of Neural Network.\n",
    "model.save_weights('NN_model_weights.h5')"
   ]
  }
 ],
 "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": "3.10.9 (tags/v3.10.9:1dd9be6, Dec  6 2022, 20:01:21) [MSC v.1934 64 bit (AMD64)]"
  },
  "vscode": {
   "interpreter": {
    "hash": "9781c75f8ebb8db568c34abc6fa56aff5b950f24b8cfd6fe8af6829c67f4ac57"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}