NeuralNetwork.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "p1b9IqDY5_Ha"
   },
   "source": [
    "## Data Preproccessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "Y6XCq8dB2OUI"
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import tensorflow as tf\n",
    "from keras.models import load_model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### how to run and import preproccessed data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 5846 rows and 3 columns\n",
      "ham     4510\n",
      "spam    1336\n",
      "Name: Category, dtype: int64\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                             Message  Label\n",
      "0  go jurong point crazi avail bugi n great world...      0\n",
      "1                              ok lar joke wif u oni      0\n",
      "3                u dun say earli hor u c alreadi say      0\n",
      "4               nah think goe usf live around though      0\n",
      "6      even brother like speak treat like aid patent      0\n",
      "X_train: (4676,)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from utils import get_cleaned_data, generate_model_output\n",
    "X_train, X_test, y_train, y_test, vocab_size, sentence_len = get_cleaned_data()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "4qfEK7JgB6zl"
   },
   "source": [
    "### Neural Network with SGD optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "XVVIVDnmB8yq",
    "outputId": "2711cf75-375a-480f-e6e5-873032419fcf"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_3\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding_3 (Embedding)     (None, 200, 32)           320000    \n",
      "                                                                 \n",
      " bidirectional_3 (Bidirectio  (None, 128)              49664     \n",
      " nal)                                                            \n",
      "                                                                 \n",
      " dense_6 (Dense)             (None, 16)                2064      \n",
      "                                                                 \n",
      " dropout_3 (Dropout)         (None, 16)                0         \n",
      "                                                                 \n",
      " dense_7 (Dense)             (None, 1)                 17        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 371,745\n",
      "Trainable params: 371,745\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\cpere\\anaconda3\\lib\\site-packages\\keras\\optimizers\\optimizer_v2\\gradient_descent.py:114: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
      "  super().__init__(name, **kwargs)\n"
     ]
    }
   ],
   "source": [
    "# model inspired from https://towardsdatascience.com/spam-detection-in-emails-de0398ea3b48\n",
    "from keras.layers import Dense, Input, LSTM, Embedding, Dropout, Activation\n",
    "from keras.layers import Bidirectional\n",
    "from keras.optimizers import SGD\n",
    "\n",
    "LR = .01\n",
    "MOMENTUM = .05\n",
    "#size of the output vector from each layer\n",
    "embedding_vector_length = 32\n",
    "\n",
    "lstm = tf.keras.Sequential()\n",
    "\n",
    "lstm.add(Embedding(vocab_size, embedding_vector_length, input_length=sentence_len)) #embedding layer to vectorize\n",
    "lstm.add(Bidirectional(tf.keras.layers.LSTM(64)))\n",
    "lstm.add(Dense(16, activation='relu'))\n",
    "lstm.add(Dropout(0.1)) #prevent overfitting\n",
    "lstm.add(Dense(1, activation='sigmoid')) # normalize output\n",
    "\n",
    "lstm.compile(loss='binary_crossentropy', optimizer= SGD(lr=LR, momentum=MOMENTUM), metrics=['accuracy'] )\n",
    "\n",
    "print(lstm.summary())\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "WEVi27fSSO4X",
    "outputId": "b6c6433a-21db-46d7-8081-44766d721222"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/17\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.6928 - accuracy: 0.5607 - val_loss: 0.6912 - val_accuracy: 0.7581\n",
      "Epoch 2/17\n",
      "15/15 [==============================] - 30s 2s/step - loss: 0.6924 - accuracy: 0.5531 - val_loss: 0.6919 - val_accuracy: 0.6915\n",
      "Epoch 3/17\n",
      "15/15 [==============================] - 32s 2s/step - loss: 0.6921 - accuracy: 0.6123 - val_loss: 0.6906 - val_accuracy: 0.7786\n",
      "Epoch 4/17\n",
      "15/15 [==============================] - 32s 2s/step - loss: 0.6919 - accuracy: 0.5904 - val_loss: 0.6896 - val_accuracy: 0.7846\n",
      "Epoch 5/17\n",
      "15/15 [==============================] - 32s 2s/step - loss: 0.6915 - accuracy: 0.5428 - val_loss: 0.6885 - val_accuracy: 0.7872\n",
      "Epoch 6/17\n",
      "15/15 [==============================] - 33s 2s/step - loss: 0.6912 - accuracy: 0.5539 - val_loss: 0.6878 - val_accuracy: 0.7846\n",
      "Epoch 7/17\n",
      "15/15 [==============================] - 33s 2s/step - loss: 0.6910 - accuracy: 0.5361 - val_loss: 0.6869 - val_accuracy: 0.7803\n",
      "Epoch 8/17\n",
      "15/15 [==============================] - 33s 2s/step - loss: 0.6908 - accuracy: 0.5250 - val_loss: 0.6873 - val_accuracy: 0.7915\n",
      "Epoch 9/17\n",
      "15/15 [==============================] - 32s 2s/step - loss: 0.6906 - accuracy: 0.5747 - val_loss: 0.6880 - val_accuracy: 0.8162\n",
      "Epoch 10/17\n",
      "15/15 [==============================] - 32s 2s/step - loss: 0.6904 - accuracy: 0.6092 - val_loss: 0.6884 - val_accuracy: 0.8231\n",
      "Epoch 11/17\n",
      "15/15 [==============================] - 32s 2s/step - loss: 0.6902 - accuracy: 0.6770 - val_loss: 0.6877 - val_accuracy: 0.8171\n",
      "Epoch 12/17\n",
      "15/15 [==============================] - 32s 2s/step - loss: 0.6900 - accuracy: 0.6902 - val_loss: 0.6862 - val_accuracy: 0.8051\n",
      "Epoch 13/17\n",
      "15/15 [==============================] - 32s 2s/step - loss: 0.6896 - accuracy: 0.6063 - val_loss: 0.6848 - val_accuracy: 0.7897\n",
      "Epoch 14/17\n",
      "15/15 [==============================] - 32s 2s/step - loss: 0.6895 - accuracy: 0.5390 - val_loss: 0.6846 - val_accuracy: 0.7932\n",
      "Epoch 15/17\n",
      "15/15 [==============================] - 33s 2s/step - loss: 0.6892 - accuracy: 0.5714 - val_loss: 0.6832 - val_accuracy: 0.7821\n",
      "Epoch 16/17\n",
      "15/15 [==============================] - 33s 2s/step - loss: 0.6888 - accuracy: 0.5460 - val_loss: 0.6829 - val_accuracy: 0.7855\n",
      "Epoch 17/17\n",
      "15/15 [==============================] - 32s 2s/step - loss: 0.6886 - accuracy: 0.5454 - val_loss: 0.6830 - val_accuracy: 0.7966\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x1cd0b86e490>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "BATCH_SIZE = 512\n",
    "EPOCHS = 17\n",
    "\n",
    "lstm.fit(X_train, y_train, batch_size=BATCH_SIZE, epochs=EPOCHS, validation_data=(X_test, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 312
    },
    "id": "-qKa35QQTzti",
    "outputId": "67dce3f2-5246-4592-af1f-cfa3bde4e9d4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dict_keys(['loss', 'accuracy', 'val_loss', 'val_accuracy'])\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot model accuracy\n",
    "print(lstm.history.history.keys())\n",
    "plt.plot(lstm.history.history['accuracy'])\n",
    "plt.plot(lstm.history.history['val_accuracy'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'test'], loc='upper left')\n",
    "plt.grid()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "37/37 [==============================] - 2s 39ms/step\n"
     ]
    }
   ],
   "source": [
    "# test our model\n",
    "y_pred=lstm.predict(X_test)\n",
    "y_pred=(y_pred>0.5) # assign true(spam) if > .5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification Report\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.80      0.99      0.88       911\n",
      "           1       0.72      0.13      0.22       259\n",
      "\n",
      "    accuracy                           0.80      1170\n",
      "   macro avg       0.76      0.56      0.55      1170\n",
      "weighted avg       0.78      0.80      0.74      1170\n",
      " \n",
      "\n",
      "Accuracy Score: 0.7965811965811965\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "generate_model_output(y_test,y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "5DUdikL3Zqwc"
   },
   "source": [
    "### Neural Network with adam optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "A8yEvjGqZy6V",
    "outputId": "96e759d8-e43d-47c3-ed68-c52486d5b304"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding_2 (Embedding)     (None, 200, 32)           320000    \n",
      "                                                                 \n",
      " bidirectional_2 (Bidirectio  (None, 128)              49664     \n",
      " nal)                                                            \n",
      "                                                                 \n",
      " dense_4 (Dense)             (None, 16)                2064      \n",
      "                                                                 \n",
      " dropout_2 (Dropout)         (None, 16)                0         \n",
      "                                                                 \n",
      " dense_5 (Dense)             (None, 1)                 17        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 371,745\n",
      "Trainable params: 371,745\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "from keras.layers import Dense, Embedding, Dropout\n",
    "from keras.layers import Bidirectional\n",
    "\n",
    "#size of the output vector from each layer\n",
    "embedding_vector_length = 32\n",
    "\n",
    "lstm = tf.keras.Sequential()\n",
    "\n",
    "lstm.add(Embedding(vocab_size, embedding_vector_length, input_length=sentence_len)) #embedding layer to vectorize\n",
    "lstm.add(Bidirectional(tf.keras.layers.LSTM(64)))\n",
    "lstm.add(Dense(16, activation='relu'))\n",
    "lstm.add(Dropout(0.1)) #prevent overfitting\n",
    "lstm.add(Dense(1, activation='sigmoid')) # normalize output\n",
    "\n",
    "lstm.compile(loss='binary_crossentropy', optimizer= 'adam', metrics=['accuracy'] )\n",
    "\n",
    "print(lstm.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "elVfAU0-Z5OF",
    "outputId": "92c12d4d-df6c-40f8-ae84-0b0f499dcff8"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/26\n",
      "15/15 [==============================] - 33s 2s/step - loss: 0.6762 - accuracy: 0.7380 - val_loss: 0.5647 - val_accuracy: 0.8803\n",
      "Epoch 2/26\n",
      "15/15 [==============================] - 26s 2s/step - loss: 0.5086 - accuracy: 0.8145 - val_loss: 0.2526 - val_accuracy: 0.9197\n",
      "Epoch 3/26\n",
      "15/15 [==============================] - 28s 2s/step - loss: 0.3037 - accuracy: 0.8844 - val_loss: 0.2277 - val_accuracy: 0.9137\n",
      "Epoch 4/26\n",
      "15/15 [==============================] - 28s 2s/step - loss: 0.2155 - accuracy: 0.9278 - val_loss: 0.1397 - val_accuracy: 0.9573\n",
      "Epoch 5/26\n",
      "15/15 [==============================] - 28s 2s/step - loss: 0.1432 - accuracy: 0.9542 - val_loss: 0.0900 - val_accuracy: 0.9709\n",
      "Epoch 6/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0990 - accuracy: 0.9686 - val_loss: 0.0503 - val_accuracy: 0.9863\n",
      "Epoch 7/26\n",
      "15/15 [==============================] - 28s 2s/step - loss: 0.0902 - accuracy: 0.9705 - val_loss: 0.1652 - val_accuracy: 0.9513\n",
      "Epoch 8/26\n",
      "15/15 [==============================] - 28s 2s/step - loss: 0.0677 - accuracy: 0.9792 - val_loss: 0.0942 - val_accuracy: 0.9701\n",
      "Epoch 9/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0522 - accuracy: 0.9849 - val_loss: 0.0838 - val_accuracy: 0.9778\n",
      "Epoch 10/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0386 - accuracy: 0.9901 - val_loss: 0.1047 - val_accuracy: 0.9726\n",
      "Epoch 11/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0315 - accuracy: 0.9922 - val_loss: 0.0903 - val_accuracy: 0.9761\n",
      "Epoch 12/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0298 - accuracy: 0.9915 - val_loss: 0.0936 - val_accuracy: 0.9786\n",
      "Epoch 13/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0243 - accuracy: 0.9936 - val_loss: 0.1042 - val_accuracy: 0.9769\n",
      "Epoch 14/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0223 - accuracy: 0.9940 - val_loss: 0.1052 - val_accuracy: 0.9769\n",
      "Epoch 15/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0201 - accuracy: 0.9947 - val_loss: 0.0592 - val_accuracy: 0.9872\n",
      "Epoch 16/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0358 - accuracy: 0.9912 - val_loss: 0.0644 - val_accuracy: 0.9821\n",
      "Epoch 17/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0242 - accuracy: 0.9946 - val_loss: 0.0814 - val_accuracy: 0.9803\n",
      "Epoch 18/26\n",
      "15/15 [==============================] - 32s 2s/step - loss: 0.0197 - accuracy: 0.9946 - val_loss: 0.0994 - val_accuracy: 0.9761\n",
      "Epoch 19/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0161 - accuracy: 0.9960 - val_loss: 0.1034 - val_accuracy: 0.9744\n",
      "Epoch 20/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0124 - accuracy: 0.9971 - val_loss: 0.1189 - val_accuracy: 0.9709\n",
      "Epoch 21/26\n",
      "15/15 [==============================] - 30s 2s/step - loss: 0.0110 - accuracy: 0.9974 - val_loss: 0.1022 - val_accuracy: 0.9795\n",
      "Epoch 22/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0094 - accuracy: 0.9978 - val_loss: 0.1196 - val_accuracy: 0.9761\n",
      "Epoch 23/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0084 - accuracy: 0.9979 - val_loss: 0.1280 - val_accuracy: 0.9752\n",
      "Epoch 24/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0071 - accuracy: 0.9982 - val_loss: 0.1457 - val_accuracy: 0.9718\n",
      "Epoch 25/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0063 - accuracy: 0.9986 - val_loss: 0.1373 - val_accuracy: 0.9744\n",
      "Epoch 26/26\n",
      "15/15 [==============================] - 29s 2s/step - loss: 0.0057 - accuracy: 0.9990 - val_loss: 0.1307 - val_accuracy: 0.9761\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x1cd0a70be80>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "BATCH_SIZE = 512\n",
    "EPOCHS = 26\n",
    "\n",
    "lstm.fit(X_train, y_train, batch_size=BATCH_SIZE, epochs=EPOCHS, validation_data=(X_test, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "zEjzUH93c2Cc",
    "outputId": "69fbb5a2-10d1-4df2-a2f2-35cd65852c2c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model Saved!\n",
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding_2 (Embedding)     (None, 200, 32)           320000    \n",
      "                                                                 \n",
      " bidirectional_2 (Bidirectio  (None, 128)              49664     \n",
      " nal)                                                            \n",
      "                                                                 \n",
      " dense_4 (Dense)             (None, 16)                2064      \n",
      "                                                                 \n",
      " dropout_2 (Dropout)         (None, 16)                0         \n",
      "                                                                 \n",
      " dense_5 (Dense)             (None, 1)                 17        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 371,745\n",
      "Trainable params: 371,745\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# save model \n",
    "lstm.save('lstmModel.h5')\n",
    "print('Model Saved!')\n",
    " \n",
    "# load model\n",
    "savedModel=load_model('lstmModel.h5')\n",
    "savedModel.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 312
    },
    "id": "n1q76FDbZ8N1",
    "outputId": "abed4cb7-a345-4a5c-950d-8002eedcd0f5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dict_keys(['loss', 'accuracy', 'val_loss', 'val_accuracy'])\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot model accuracy\n",
    "print(lstm.history.history.keys())\n",
    "plt.plot(lstm.history.history['accuracy'])\n",
    "plt.plot(lstm.history.history['val_accuracy'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'test'], loc='upper left')\n",
    "plt.grid()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "ZeIt5kz2YpG1",
    "outputId": "e5ef9123-b1a1-4340-b953-cdc613cabc8f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "37/37 [==============================] - 2s 30ms/step\n"
     ]
    }
   ],
   "source": [
    "# test our model\n",
    "y_pred=lstm.predict(X_test)\n",
    "y_pred=(y_pred>0.5) # assign true(spam) if > .5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "xCG_3cgmYJVk",
    "outputId": "316714f8-1009-40ce-c868-f9648bf01158"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification Report\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      0.97      0.98       911\n",
      "           1       0.91      1.00      0.95       259\n",
      "\n",
      "    accuracy                           0.98      1170\n",
      "   macro avg       0.95      0.98      0.97      1170\n",
      "weighted avg       0.98      0.98      0.98      1170\n",
      " \n",
      "\n",
      "Accuracy Score: 0.976068376068376\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "generate_model_output(y_test,y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "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": "3.9.10 (main, Jan 15 2022, 11:40:53) \n[Clang 13.0.0 (clang-1300.0.29.3)]"
  },
  "vscode": {
   "interpreter": {
    "hash": "949777d72b0d2535278d3dc13498b2535136f6dfe0678499012e853ee9abcab1"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}