Sentiment Analysis with Deep Learning - Tensorflow.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "f989b19f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# !pip install git+ssh://git@github.com/seroetr/preprocess_seroetr.git #download necessary packages\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "9490ceac",
   "metadata": {},
   "outputs": [],
   "source": [
    "# !pip install -U spacy #download necessary packages\n",
    "# !python -m spacy download en_core_web_lg #download necessary packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "137c2724",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import spacy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "d875b9b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "import preprocess_seroetr as ps\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import confusion_matrix, classification_report"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "17570704",
   "metadata": {},
   "source": [
    "#### Data preprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "874c48ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "df=pd.read_csv(\"imdb_reviews.txt\",sep=\"\\t\",header=None)\n",
    "df.columns=[\"reviews\",\"sentiment\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "a3a4d5f3",
   "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>reviews</th>\n",
       "      <th>sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>A very, very, very slow-moving, aimless movie ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Not sure who was more lost - the flat characte...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Attempting artiness with black &amp; white and cle...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Very little music or anything to speak of.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>The best scene in the movie was when Gerardo i...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             reviews  sentiment\n",
       "0  A very, very, very slow-moving, aimless movie ...          0\n",
       "1  Not sure who was more lost - the flat characte...          0\n",
       "2  Attempting artiness with black & white and cle...          0\n",
       "3       Very little music or anything to speak of.            0\n",
       "4  The best scene in the movie was when Gerardo i...          1"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "86f311ea",
   "metadata": {},
   "outputs": [],
   "source": [
    "def clean_data(x):\n",
    "    x = str(x).lower()\n",
    "    x = ps.get_cont_to_exp(x)\n",
    "    x = ps.remove_emails(x)\n",
    "    x = ps.remove_html_tags(x)\n",
    "    x = ps.remove_urls(x)\n",
    "    x = ps.remove_special_chars(x)\n",
    "    x = ps.remove_accented_chars(x)\n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "e63ab934",
   "metadata": {},
   "outputs": [],
   "source": [
    "x=\"that's a really scary movie!\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "dbcd2432",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "that's a really scary movie!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'that is a really scary movie'"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(x)\n",
    "clean_data(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "d827ac63",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\developer\\appdata\\local\\programs\\python\\python38\\lib\\site-packages\\bs4\\__init__.py:435: MarkupResemblesLocatorWarning: The input looks more like a filename than markup. You may want to open this file and pass the filehandle into Beautiful Soup.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "df[\"reviews\"]=df[\"reviews\"].apply(lambda x: clean_data(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "55bb669c",
   "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>reviews</th>\n",
       "      <th>sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>a very very very slowmoving aimless movie abou...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>not sure who was more lost the flat characters...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>attempting artiness with black white and cleve...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>very little music or anything to speak of</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>the best scene in the movie was when gerardo i...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             reviews  sentiment\n",
       "0  a very very very slowmoving aimless movie abou...          0\n",
       "1  not sure who was more lost the flat characters...          0\n",
       "2  attempting artiness with black white and cleve...          0\n",
       "3          very little music or anything to speak of          0\n",
       "4  the best scene in the movie was when gerardo i...          1"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "086fe14f",
   "metadata": {},
   "outputs": [],
   "source": [
    "nlp=spacy.load(\"en_core_web_lg\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "0ef3d7b6",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_vec(x):\n",
    "    doc=nlp(x)\n",
    "    vec=doc.vector\n",
    "    return vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "9636010b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(300,)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_vec(\"how are you\").shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "1eb765f7",
   "metadata": {},
   "outputs": [],
   "source": [
    "df[\"vec\"]=df[\"reviews\"].apply(lambda x: get_vec(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "4f0292ab",
   "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>reviews</th>\n",
       "      <th>sentiment</th>\n",
       "      <th>vec</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>a very very very slowmoving aimless movie abou...</td>\n",
       "      <td>0</td>\n",
       "      <td>[-1.7658583, 1.6712562, -2.762896, -1.448437, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>not sure who was more lost the flat characters...</td>\n",
       "      <td>0</td>\n",
       "      <td>[-3.794328, 0.70210844, -2.5325108, 0.14926393...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>attempting artiness with black white and cleve...</td>\n",
       "      <td>0</td>\n",
       "      <td>[-2.9543035, -0.07106956, -2.2745807, -0.66580...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>very little music or anything to speak of</td>\n",
       "      <td>0</td>\n",
       "      <td>[-1.9868913, 2.2468886, -4.532146, -2.550426, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>the best scene in the movie was when gerardo i...</td>\n",
       "      <td>1</td>\n",
       "      <td>[-1.0748299, 2.742419, -2.5380828, -1.3163186,...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             reviews  sentiment  \\\n",
       "0  a very very very slowmoving aimless movie abou...          0   \n",
       "1  not sure who was more lost the flat characters...          0   \n",
       "2  attempting artiness with black white and cleve...          0   \n",
       "3          very little music or anything to speak of          0   \n",
       "4  the best scene in the movie was when gerardo i...          1   \n",
       "\n",
       "                                                 vec  \n",
       "0  [-1.7658583, 1.6712562, -2.762896, -1.448437, ...  \n",
       "1  [-3.794328, 0.70210844, -2.5325108, 0.14926393...  \n",
       "2  [-2.9543035, -0.07106956, -2.2745807, -0.66580...  \n",
       "3  [-1.9868913, 2.2468886, -4.532146, -2.550426, ...  \n",
       "4  [-1.0748299, 2.742419, -2.5380828, -1.3163186,...  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "8c27ac25",
   "metadata": {},
   "outputs": [],
   "source": [
    "y=df[\"sentiment\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "70165656",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df['vec'].to_numpy()\n",
    "X = X.reshape(-1, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "b2c1245d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((748, 3), (748, 1))"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape, X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "29d3d4f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.concatenate(np.concatenate(X, axis = 0), axis = 0).reshape(-1, 300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "3db3c4d0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((748, 3), (748, 300))"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape, X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "5ed38bab",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 0, stratify = y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "b1af619b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((598, 300), (150, 300))"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape, X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "4d9749db",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-11.106075"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.min(X_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "39c08d34",
   "metadata": {},
   "source": [
    "#### Feature scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "4ecb0f07",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "14d42e50",
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler=MinMaxScaler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "98916096",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train=scaler.fit_transform(X_train) # fitting values between 0-1\n",
    "X_test=scaler.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "68f6d244",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.0, 1.0000001)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.min(X_train), np.max(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "6f606ce2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "300"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape[1]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a34434d6",
   "metadata": {},
   "source": [
    "#### Artificial Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "740cbac3",
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import Sequential\n",
    "from tensorflow.keras.layers import Flatten, Dense\n",
    "from tensorflow.keras.optimizers import Adam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "721e533b",
   "metadata": {},
   "outputs": [],
   "source": [
    "model=Sequential()\n",
    "model.add(Dense(128,activation=\"relu\",input_dim=X.shape[1])) # input layer\n",
    "model.add(Dense(128,activation=\"relu\")) # hidden layer\n",
    "model.add(Dense(128,activation=\"relu\")) # hidden layer\n",
    "model.add(Dense(1,activation=\"sigmoid\")) # output layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "42b5b79f",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\developer\\appdata\\local\\programs\\python\\python38\\lib\\site-packages\\keras\\optimizers\\optimizer_v2\\adam.py:110: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
      "  super(Adam, self).__init__(name, **kwargs)\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer=Adam(lr=0.00005),loss=\"binary_crossentropy\",metrics=[\"accuracy\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "bddec3dc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "75/75 [==============================] - 1s 3ms/step - loss: 0.6895 - accuracy: 0.5184\n",
      "Epoch 2/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.6769 - accuracy: 0.6104\n",
      "Epoch 3/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.6670 - accuracy: 0.6371\n",
      "Epoch 4/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.6605 - accuracy: 0.6388\n",
      "Epoch 5/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.6513 - accuracy: 0.6572\n",
      "Epoch 6/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.6414 - accuracy: 0.6839\n",
      "Epoch 7/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.6347 - accuracy: 0.6689\n",
      "Epoch 8/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.6195 - accuracy: 0.6957\n",
      "Epoch 9/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.6150 - accuracy: 0.6856\n",
      "Epoch 10/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5989 - accuracy: 0.7090\n",
      "Epoch 11/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5878 - accuracy: 0.7258\n",
      "Epoch 12/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5770 - accuracy: 0.7425\n",
      "Epoch 13/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5693 - accuracy: 0.7057\n",
      "Epoch 14/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5627 - accuracy: 0.7324\n",
      "Epoch 15/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5548 - accuracy: 0.7291\n",
      "Epoch 16/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5534 - accuracy: 0.7341\n",
      "Epoch 17/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5401 - accuracy: 0.7575\n",
      "Epoch 18/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5324 - accuracy: 0.7492\n",
      "Epoch 19/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5279 - accuracy: 0.7625\n",
      "Epoch 20/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5188 - accuracy: 0.7625\n",
      "Epoch 21/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5205 - accuracy: 0.7559\n",
      "Epoch 22/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5126 - accuracy: 0.7625\n",
      "Epoch 23/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5010 - accuracy: 0.7843\n",
      "Epoch 24/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.5036 - accuracy: 0.7742\n",
      "Epoch 25/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4884 - accuracy: 0.7759\n",
      "Epoch 26/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4823 - accuracy: 0.7943\n",
      "Epoch 27/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4770 - accuracy: 0.7876\n",
      "Epoch 28/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4744 - accuracy: 0.7893\n",
      "Epoch 29/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4720 - accuracy: 0.8027\n",
      "Epoch 30/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.4728 - accuracy: 0.7793\n",
      "Epoch 31/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4538 - accuracy: 0.8043\n",
      "Epoch 32/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4590 - accuracy: 0.8010\n",
      "Epoch 33/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4572 - accuracy: 0.7993\n",
      "Epoch 34/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4441 - accuracy: 0.8077\n",
      "Epoch 35/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4505 - accuracy: 0.7943\n",
      "Epoch 36/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4443 - accuracy: 0.8010\n",
      "Epoch 37/200\n",
      "75/75 [==============================] - 0s 5ms/step - loss: 0.4285 - accuracy: 0.8244\n",
      "Epoch 38/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.4294 - accuracy: 0.8227\n",
      "Epoch 39/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4341 - accuracy: 0.8227\n",
      "Epoch 40/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.4351 - accuracy: 0.8127\n",
      "Epoch 41/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4215 - accuracy: 0.8127\n",
      "Epoch 42/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.4165 - accuracy: 0.8211\n",
      "Epoch 43/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4159 - accuracy: 0.8161\n",
      "Epoch 44/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4156 - accuracy: 0.8060\n",
      "Epoch 45/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4034 - accuracy: 0.8244\n",
      "Epoch 46/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.4079 - accuracy: 0.8278\n",
      "Epoch 47/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3988 - accuracy: 0.8328\n",
      "Epoch 48/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3987 - accuracy: 0.8378\n",
      "Epoch 49/200\n",
      "75/75 [==============================] - 0s 5ms/step - loss: 0.3945 - accuracy: 0.8177\n",
      "Epoch 50/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.4042 - accuracy: 0.8211\n",
      "Epoch 51/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3832 - accuracy: 0.8428\n",
      "Epoch 52/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.3811 - accuracy: 0.8495\n",
      "Epoch 53/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3893 - accuracy: 0.8294\n",
      "Epoch 54/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.3821 - accuracy: 0.8361\n",
      "Epoch 55/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3850 - accuracy: 0.8411\n",
      "Epoch 56/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3743 - accuracy: 0.8411\n",
      "Epoch 57/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.3666 - accuracy: 0.8462\n",
      "Epoch 58/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3647 - accuracy: 0.8428\n",
      "Epoch 59/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.3689 - accuracy: 0.8244\n",
      "Epoch 60/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3771 - accuracy: 0.8411\n",
      "Epoch 61/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3602 - accuracy: 0.8445\n",
      "Epoch 62/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.3534 - accuracy: 0.8562\n",
      "Epoch 63/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.3648 - accuracy: 0.8462\n",
      "Epoch 64/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.3491 - accuracy: 0.8595\n",
      "Epoch 65/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.3497 - accuracy: 0.8478\n",
      "Epoch 66/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.3576 - accuracy: 0.8445\n",
      "Epoch 67/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.3511 - accuracy: 0.8445\n",
      "Epoch 68/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3427 - accuracy: 0.8579\n",
      "Epoch 69/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3438 - accuracy: 0.8462\n",
      "Epoch 70/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.3432 - accuracy: 0.8528\n",
      "Epoch 71/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3309 - accuracy: 0.8612\n",
      "Epoch 72/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3338 - accuracy: 0.8612\n",
      "Epoch 73/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.3283 - accuracy: 0.8629\n",
      "Epoch 74/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.3317 - accuracy: 0.8462\n",
      "Epoch 75/200\n",
      "75/75 [==============================] - 0s 5ms/step - loss: 0.3337 - accuracy: 0.8645\n",
      "Epoch 76/200\n",
      "75/75 [==============================] - 0s 5ms/step - loss: 0.3215 - accuracy: 0.8712\n",
      "Epoch 77/200\n",
      "75/75 [==============================] - 0s 5ms/step - loss: 0.3229 - accuracy: 0.8629\n",
      "Epoch 78/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.3257 - accuracy: 0.8645\n",
      "Epoch 79/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.3203 - accuracy: 0.8763\n",
      "Epoch 80/200\n",
      "75/75 [==============================] - 0s 5ms/step - loss: 0.3174 - accuracy: 0.8612\n",
      "Epoch 81/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.3097 - accuracy: 0.8729\n",
      "Epoch 82/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.3208 - accuracy: 0.8729\n",
      "Epoch 83/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3138 - accuracy: 0.8679\n",
      "Epoch 84/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.3036 - accuracy: 0.8629\n",
      "Epoch 85/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.3084 - accuracy: 0.8729\n",
      "Epoch 86/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3084 - accuracy: 0.8796\n",
      "Epoch 87/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.3078 - accuracy: 0.8629\n",
      "Epoch 88/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2968 - accuracy: 0.8746\n",
      "Epoch 89/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2944 - accuracy: 0.8813\n",
      "Epoch 90/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2973 - accuracy: 0.8712\n",
      "Epoch 91/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.3043 - accuracy: 0.8863\n",
      "Epoch 92/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2942 - accuracy: 0.8679\n",
      "Epoch 93/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.2913 - accuracy: 0.8863\n",
      "Epoch 94/200\n",
      "75/75 [==============================] - 0s 5ms/step - loss: 0.2900 - accuracy: 0.8829\n",
      "Epoch 95/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2828 - accuracy: 0.8846\n",
      "Epoch 96/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2891 - accuracy: 0.8813\n",
      "Epoch 97/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2803 - accuracy: 0.8930\n",
      "Epoch 98/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2781 - accuracy: 0.8863\n",
      "Epoch 99/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2768 - accuracy: 0.8896\n",
      "Epoch 100/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2830 - accuracy: 0.8913\n",
      "Epoch 101/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2802 - accuracy: 0.8880\n",
      "Epoch 102/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2842 - accuracy: 0.8896\n",
      "Epoch 103/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2703 - accuracy: 0.8846\n",
      "Epoch 104/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2633 - accuracy: 0.8930\n",
      "Epoch 105/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2806 - accuracy: 0.8829\n",
      "Epoch 106/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2553 - accuracy: 0.8913\n",
      "Epoch 107/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2679 - accuracy: 0.8880\n",
      "Epoch 108/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2694 - accuracy: 0.8896\n",
      "Epoch 109/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2689 - accuracy: 0.8729\n",
      "Epoch 110/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2553 - accuracy: 0.8946\n",
      "Epoch 111/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2536 - accuracy: 0.8980\n",
      "Epoch 112/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2465 - accuracy: 0.8997\n",
      "Epoch 113/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2603 - accuracy: 0.8946\n",
      "Epoch 114/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2417 - accuracy: 0.9047\n",
      "Epoch 115/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2691 - accuracy: 0.8863\n",
      "Epoch 116/200\n",
      "75/75 [==============================] - 0s 5ms/step - loss: 0.2405 - accuracy: 0.8997\n",
      "Epoch 117/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2476 - accuracy: 0.9047\n",
      "Epoch 118/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2391 - accuracy: 0.8963\n",
      "Epoch 119/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2313 - accuracy: 0.9097\n",
      "Epoch 120/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2427 - accuracy: 0.8963\n",
      "Epoch 121/200\n",
      "75/75 [==============================] - 0s 5ms/step - loss: 0.2317 - accuracy: 0.9080\n",
      "Epoch 122/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2275 - accuracy: 0.9047\n",
      "Epoch 123/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2278 - accuracy: 0.9114\n",
      "Epoch 124/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2354 - accuracy: 0.9097\n",
      "Epoch 125/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2331 - accuracy: 0.9030\n",
      "Epoch 126/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.2254 - accuracy: 0.9013\n",
      "Epoch 127/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2279 - accuracy: 0.9181\n",
      "Epoch 128/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2256 - accuracy: 0.9114\n",
      "Epoch 129/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2333 - accuracy: 0.8913\n",
      "Epoch 130/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2446 - accuracy: 0.8913\n",
      "Epoch 131/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2227 - accuracy: 0.9047\n",
      "Epoch 132/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2171 - accuracy: 0.9147\n",
      "Epoch 133/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2164 - accuracy: 0.9181\n",
      "Epoch 134/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2128 - accuracy: 0.9130\n",
      "Epoch 135/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2243 - accuracy: 0.8980\n",
      "Epoch 136/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2122 - accuracy: 0.9214\n",
      "Epoch 137/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2111 - accuracy: 0.9247\n",
      "Epoch 138/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2129 - accuracy: 0.9130\n",
      "Epoch 139/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2028 - accuracy: 0.9264\n",
      "Epoch 140/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2045 - accuracy: 0.9281\n",
      "Epoch 141/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2033 - accuracy: 0.9247\n",
      "Epoch 142/200\n",
      "75/75 [==============================] - 0s 6ms/step - loss: 0.2052 - accuracy: 0.9181\n",
      "Epoch 143/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.2154 - accuracy: 0.9197\n",
      "Epoch 144/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.1997 - accuracy: 0.9247\n",
      "Epoch 145/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.1956 - accuracy: 0.9298\n",
      "Epoch 146/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1987 - accuracy: 0.9214\n",
      "Epoch 147/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1920 - accuracy: 0.9281\n",
      "Epoch 148/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1964 - accuracy: 0.9231\n",
      "Epoch 149/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1865 - accuracy: 0.9247\n",
      "Epoch 150/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1843 - accuracy: 0.9348\n",
      "Epoch 151/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1917 - accuracy: 0.9247\n",
      "Epoch 152/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.1832 - accuracy: 0.9365\n",
      "Epoch 153/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1851 - accuracy: 0.9247\n",
      "Epoch 154/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1903 - accuracy: 0.9231\n",
      "Epoch 155/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1781 - accuracy: 0.9415\n",
      "Epoch 156/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1795 - accuracy: 0.9331\n",
      "Epoch 157/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1762 - accuracy: 0.9365\n",
      "Epoch 158/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1797 - accuracy: 0.9415\n",
      "Epoch 159/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.1730 - accuracy: 0.9331\n",
      "Epoch 160/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1666 - accuracy: 0.9348\n",
      "Epoch 161/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1659 - accuracy: 0.9398\n",
      "Epoch 162/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1730 - accuracy: 0.9381\n",
      "Epoch 163/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1732 - accuracy: 0.9398\n",
      "Epoch 164/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1695 - accuracy: 0.9381\n",
      "Epoch 165/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1661 - accuracy: 0.9381\n",
      "Epoch 166/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1651 - accuracy: 0.9415\n",
      "Epoch 167/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1592 - accuracy: 0.9498\n",
      "Epoch 168/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1630 - accuracy: 0.9331\n",
      "Epoch 169/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1581 - accuracy: 0.9348\n",
      "Epoch 170/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1729 - accuracy: 0.9314\n",
      "Epoch 171/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1552 - accuracy: 0.9465\n",
      "Epoch 172/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1625 - accuracy: 0.9365\n",
      "Epoch 173/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1479 - accuracy: 0.9515\n",
      "Epoch 174/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.1498 - accuracy: 0.9482\n",
      "Epoch 175/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1443 - accuracy: 0.9532\n",
      "Epoch 176/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1452 - accuracy: 0.9565\n",
      "Epoch 177/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1583 - accuracy: 0.9515\n",
      "Epoch 178/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1642 - accuracy: 0.9431\n",
      "Epoch 179/200\n",
      "75/75 [==============================] - 0s 4ms/step - loss: 0.1423 - accuracy: 0.9532\n",
      "Epoch 180/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1466 - accuracy: 0.9431\n",
      "Epoch 181/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1447 - accuracy: 0.9498\n",
      "Epoch 182/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1366 - accuracy: 0.9565\n",
      "Epoch 183/200\n",
      "75/75 [==============================] - 0s 5ms/step - loss: 0.1345 - accuracy: 0.9582\n",
      "Epoch 184/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1534 - accuracy: 0.9448\n",
      "Epoch 185/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1310 - accuracy: 0.9582\n",
      "Epoch 186/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1512 - accuracy: 0.9498\n",
      "Epoch 187/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1245 - accuracy: 0.9632\n",
      "Epoch 188/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1273 - accuracy: 0.9515\n",
      "Epoch 189/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1305 - accuracy: 0.9532\n",
      "Epoch 190/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1234 - accuracy: 0.9599\n",
      "Epoch 191/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1262 - accuracy: 0.9582\n",
      "Epoch 192/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1374 - accuracy: 0.9498\n",
      "Epoch 193/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1343 - accuracy: 0.9498\n",
      "Epoch 194/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1241 - accuracy: 0.9666\n",
      "Epoch 195/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1234 - accuracy: 0.9615\n",
      "Epoch 196/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1252 - accuracy: 0.9599\n",
      "Epoch 197/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1192 - accuracy: 0.9599\n",
      "Epoch 198/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1226 - accuracy: 0.9649\n",
      "Epoch 199/200\n",
      "75/75 [==============================] - 0s 3ms/step - loss: 0.1102 - accuracy: 0.9649\n",
      "Epoch 200/200\n",
      "75/75 [==============================] - 0s 2ms/step - loss: 0.1212 - accuracy: 0.9582\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x27736fa0fd0>"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X_train, y_train, batch_size=8, epochs=200)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "61b09692",
   "metadata": {},
   "outputs": [],
   "source": [
    "# !pip install mlxtend #download necessary packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "1e4a8865",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5/5 [==============================] - 0s 2ms/step\n"
     ]
    }
   ],
   "source": [
    "y_pred=model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "3be9894f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[9.99881089e-01]\n",
      " [7.47601851e-04]\n",
      " [9.43031549e-01]\n",
      " [7.17046142e-01]\n",
      " [9.89502847e-01]\n",
      " [1.24498144e-01]\n",
      " [9.97476220e-01]\n",
      " [6.93802722e-03]\n",
      " [3.50425005e-01]\n",
      " [9.99605954e-01]\n",
      " [1.19113281e-01]\n",
      " [9.66794431e-01]\n",
      " [9.99765873e-01]\n",
      " [9.67118740e-01]\n",
      " [9.99769926e-01]\n",
      " [2.35984474e-03]\n",
      " [3.82380784e-02]\n",
      " [7.65854895e-01]\n",
      " [3.82924937e-02]\n",
      " [1.45210661e-02]\n",
      " [8.14138174e-01]\n",
      " [5.60116780e-04]\n",
      " [9.28311348e-01]\n",
      " [4.35784161e-01]\n",
      " [9.97812271e-01]\n",
      " [9.95750368e-01]\n",
      " [9.89824653e-01]\n",
      " [1.38073519e-01]\n",
      " [5.53647578e-01]\n",
      " [8.04929912e-01]\n",
      " [2.09315084e-02]\n",
      " [7.51349688e-01]\n",
      " [8.99084210e-01]\n",
      " [5.29846177e-02]\n",
      " [2.19898954e-01]\n",
      " [9.98760223e-01]\n",
      " [9.95595276e-01]\n",
      " [1.60047293e-01]\n",
      " [7.84335971e-01]\n",
      " [9.98820186e-01]\n",
      " [5.59226833e-02]\n",
      " [9.89912570e-01]\n",
      " [7.50278711e-01]\n",
      " [9.81381893e-01]\n",
      " [8.99536163e-02]\n",
      " [9.96913493e-01]\n",
      " [2.37079769e-01]\n",
      " [1.21524499e-04]\n",
      " [3.54501158e-02]\n",
      " [9.66344059e-01]\n",
      " [2.18908325e-01]\n",
      " [7.24623740e-01]\n",
      " [3.95712376e-01]\n",
      " [8.29443634e-01]\n",
      " [3.37158702e-03]\n",
      " [9.67699468e-01]\n",
      " [9.83221903e-02]\n",
      " [9.94518325e-02]\n",
      " [7.78678199e-03]\n",
      " [3.11189473e-01]\n",
      " [9.44769025e-01]\n",
      " [9.99842823e-01]\n",
      " [4.33688074e-05]\n",
      " [1.83324013e-02]\n",
      " [5.96784121e-05]\n",
      " [2.86761999e-01]\n",
      " [9.22795594e-01]\n",
      " [9.34584677e-01]\n",
      " [6.29140362e-02]\n",
      " [8.73311460e-01]\n",
      " [8.73615205e-01]\n",
      " [2.43316770e-01]\n",
      " [3.47914803e-03]\n",
      " [5.05781285e-02]\n",
      " [9.75052655e-01]\n",
      " [6.47682488e-01]\n",
      " [4.97876527e-03]\n",
      " [2.20326129e-02]\n",
      " [9.74789858e-01]\n",
      " [1.66845307e-01]\n",
      " [3.38553041e-01]\n",
      " [9.83631492e-01]\n",
      " [7.34420896e-01]\n",
      " [4.55162704e-01]\n",
      " [7.81539083e-01]\n",
      " [2.62047648e-01]\n",
      " [9.91006970e-01]\n",
      " [2.83414215e-01]\n",
      " [7.65492022e-01]\n",
      " [6.08246505e-01]\n",
      " [6.82962546e-03]\n",
      " [6.00331783e-01]\n",
      " [9.65836551e-03]\n",
      " [9.99791145e-01]\n",
      " [9.92532730e-01]\n",
      " [4.11592424e-01]\n",
      " [8.47855508e-02]\n",
      " [9.98373151e-01]\n",
      " [9.92531717e-01]\n",
      " [7.35248566e-01]\n",
      " [8.15063000e-01]\n",
      " [9.97994840e-01]\n",
      " [9.94604290e-01]\n",
      " [6.41809165e-01]\n",
      " [4.60351765e-01]\n",
      " [1.75869599e-01]\n",
      " [8.50341976e-01]\n",
      " [7.46294856e-01]\n",
      " [3.72947752e-02]\n",
      " [2.05790112e-03]\n",
      " [1.11697875e-01]\n",
      " [9.06755805e-01]\n",
      " [1.54986739e-01]\n",
      " [4.48520303e-01]\n",
      " [9.50898826e-01]\n",
      " [9.85216677e-01]\n",
      " [1.24827237e-03]\n",
      " [1.03483468e-01]\n",
      " [6.60576997e-03]\n",
      " [2.03745067e-02]\n",
      " [1.18676228e-02]\n",
      " [9.99899745e-01]\n",
      " [3.68361128e-04]\n",
      " [2.93878950e-02]\n",
      " [8.12817395e-01]\n",
      " [9.30195391e-01]\n",
      " [9.52005684e-01]\n",
      " [9.60106552e-01]\n",
      " [9.94588673e-01]\n",
      " [1.65322900e-03]\n",
      " [1.08369030e-01]\n",
      " [9.97110963e-01]\n",
      " [6.27164602e-01]\n",
      " [1.02563873e-02]\n",
      " [5.15271276e-02]\n",
      " [9.95638430e-01]\n",
      " [7.47212827e-01]\n",
      " [9.91033733e-01]\n",
      " [3.70147079e-01]\n",
      " [9.30876970e-01]\n",
      " [5.90672977e-02]\n",
      " [6.89992085e-02]\n",
      " [9.65768039e-01]\n",
      " [7.89273938e-04]\n",
      " [9.42372501e-01]\n",
      " [1.52371684e-03]\n",
      " [1.62685275e-01]\n",
      " [7.59883493e-04]\n",
      " [1.81308074e-03]\n",
      " [7.12100029e-01]]\n"
     ]
    }
   ],
   "source": [
    "print(y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "0168c061",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred=np.where(y_pred > 0.6, 1,0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "d4fc82d0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[55 18]\n",
      " [19 58]]\n"
     ]
    }
   ],
   "source": [
    "print(confusion_matrix(y_test,y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "b291eba9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(<Figure size 432x288 with 1 Axes>,\n",
       " <AxesSubplot:xlabel='predicted label', ylabel='true label'>)"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQMAAAEGCAYAAABhHPB4AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAAAP20lEQVR4nO3debhVdbnA8e8LiDigYWKmppCCpF0zBSxBHzU1HFLsatehzCFxwrLBKa+Z92o54NWuqYlKdq9mmlqOSKaFhZWiCc6GU07FUCJgmXDe+8f5oVsuHPYx1l6A38/z8LD32vus9R54nu9Za+21z47MRJK61D2ApKWDMZAEGANJhTGQBBgDSUW3ugdoFN1XyejRq+4x1Akf7bd23SOoE5577lmmT58eC3ts6YpBj16sOPiYusdQJ0y4/YS6R1AnDNlq4CIf8zBBEmAMJBXGQBJgDCQVxkASYAwkFcZAEmAMJBXGQBJgDCQVxkASYAwkFcZAEmAMJBXGQBJgDCQVxkASYAwkFcZAEmAMJBXGQBJgDCQVxkASYAwkFcZAEmAMJBXGQBJgDCQVxkASYAwkFcZAEmAMJBXGQBJgDCQVxkASYAwkFcZAEmAMJBXGQBJgDCQVxkASYAwkFcZAEmAMJBXGQBJgDCQVxkASYAwkFcZAEgDd6h5gefP4lUcw62//YN68NubOa2Po0f/DyQcO4ZBdP8K0V14D4NQxdzPu3qdrnlQAh3/hEMbedgu911qL+x98GIBJDz7IMUcfwet//zvdunXj/AsuYtDgwTVPWr1KYxARw4DvAF2ByzLzzCq3t7QY9tWrmfHq39627ILrJ3L+j++taSItyuc+fxBHHDWSLxxy4JvLTj7peE4+5VQ+OWwXbh97GyefdDw/u/OX9Q3ZIpUdJkREV+BCYBdgE2C/iNikqu1J78TQbbZljTXWeNuyiODVV18FYObMmbx/nXXqGK3lqtwzGAxMycynASLiR8CewKMVbrN2mcnNZ32GTLj81gcZc+skAI7Ycwv232lTHnjyT5z4vbt4ZfbrNU+qRTnn3PP51G6f5KQTvkZbWxu/uPueukdqiSpPIK4LPN9w/4Wy7G0iYkRETIyIifnGnArHaY1PHHsVWx/5A4Z//cccvscWDPmX9bj0pt+zyYGXsNXh3+dPM2Zz5hE71D2mOjD6kos5e9R5THnmec4edR5Hjji07pFaovZXEzJzdGYOzMyBscIqdY/zT3tpxmwApr3yGjdNeJJBA9Zh6iuv0daWZMKY2yYxcOP31zylOnLV//6A4Xt9GoB/3XsfJt737jjXU2UMXgQ+0HB/vbJsubVyjxVYdaXub97eccu+PPLsNNZe463I7Tm0P48+O72uEdWE96+zDr+6ezwAv/zFXWy0Ub+aJ2qNKs8Z3Af0i4i+tEdgX2D/CrdXu7V6rcw132z/idKtaxeuuetR7rjvGS4/YTc22+h9ZCbP/Wkmx5w/ruZJNd+Bn92PX43/JdOnT2fDPutxyjdO48KLL+W4r3yJuXPnsmKPHnz34tF1j9kSkZnVrTxiV+B82l9aHJOZZ3T0/C6rrZcrDj6msnm05P319hPqHkGdMGSrgdx//8RY2GOVXmeQmbcBt1W5DUlLRu0nECUtHYyBJMAYSCqMgSTAGEgqjIEkwBhIKoyBJMAYSCqMgSTAGEgqjIEkwBhIKoyBJMAYSCqMgSTAGEgqjIEkwBhIKoyBJMAYSCqMgSTAGEgqjIEkwBhIKoyBJMAYSCoW+VmLETELmP+prPM/qDHL7czM1SqeTVILLTIGmdmzlYNIqldThwkRMTQiDi6314yIvtWOJanVFhuDiDgVOAE4qSzqDlxZ5VCSWq+ZPYO9gD2AOQCZ+RLgIYS0nGkmBv/IzKScTIyIVaodSVIdmonBtRFxCfCeiDgM+DlwabVjSWq1Rb6aMF9mjoqInYBXgf7ANzLzjsonk9RSi41B8RCwEu2HCg9VN46kujTzasIXgHuBTwN7A7+NiEOqHkxSazWzZ3Ac8NHMnAEQEe8F7gHGVDmYpNZq5gTiDGBWw/1ZZZmk5UhH7034Srk5BfhdRNxI+zmDPYHJLZhNUgt1dJgw/8Kip8qf+W6sbhxJdenojUqntXIQSfVa7AnEiOgNHA9sCvSYvzwzd6hwLkkt1swJxKuAx4G+wGnAs8B9Fc4kqQbNxOC9mXk58EZmjs/MQwD3CqTlTDPXGbxR/n45InYDXgLWqG4kSXVoJganR8TqwFeBC4DVgC9XOpWklmvmjUq3lJszge2rHUdSXTq66OgC3vqFqP9PZn5xSQ+zeb+1GX/LcUt6tapQr0Ej6x5BnfD6E39c5GMd7RlMXPKjSFpadXTR0Q9aOYikevkhKpIAYyCpMAaSgOZ+01H/iLgzIh4u9zeLiH+vfjRJrdTMnsGltH+AyhsAmTkZ2LfKoSS1XjMxWDkz711g2dwqhpFUn2ZiMD0iNuStD1HZG3i50qkktVwz7004GhgNDIiIF4FngM9WOpWklmvmvQlPAzuWj1XrkpmzFvc1kpY9zfymo28scB+AzPyPimaSVINmDhPmNNzuAewOPFbNOJLq0sxhwrmN9yNiFDCusokk1eKdXIG4MrDekh5EUr2aOWfwEG/9XoOuQG/A8wXScqaZcwa7N9yeC/w5M73oSFrOdBiDiOgKjMvMAS2aR1JNOjxnkJnzgCciYv0WzSOpJs0cJvQCHomIe2l4mTEz96hsKkkt10wMTql8Ckm1ayYGu2bmCY0LIuIsYHw1I0mqQzPXGey0kGW7LOlBJNWro89NOBI4CvhgRExueKgnMKHqwSS1VkeHCT8ExgLfBk5sWD4rM/9S6VSSWq6jz02YSftHqu3XunEk1cXfjiwJMAaSCmMgCTAGkgpjIAkwBpIKYyAJMAaSCmMgCTAGkgpjIAkwBpIKYyAJMAaSCmMgCTAGkgpjIAkwBpIKYyAJMAaSCmMgCTAGkgpjIAlo7rMW1QlHHX4ot4+9ld691+J397d/ENVDkydx7DFHMWfObNbfYAMu+/6VrLbaajVPqvkev/U0Zs15nXltbcyd18bQA85ms/7rcsHJ+7Liiiswd14bx37rGiY+8lzdo1aqsj2DiBgTEVMj4uGqtrE0OuBzn+eGG29727KRR47gtNO/xW8nTuJTewznO+eNqmk6LcqwEd/hY/ueydADzgbgjGOHc8bosXxs3zP5z4tv4Yxjh9c7YAtUeZhwBTCswvUvlYYM3ZZea6zxtmVPTXmSIUO3BWD7HXbipp/eUMdo6oRMWG2VHgCsvupKvDxtZs0TVa+yw4TMvDsi+lS1/mXJgA9tyq0338juewznpzdcx4svPF/3SGqQmdx80Ugyk8uvn8CYGyZw3KjruPnCo/n2l/eiS5dg+4POrXvMytV+AjEiRkTExIiYOH3atLrHqcRFl1zGpaMvZtutBzFr9ixW6N697pHU4BMHn8fW+5/F8JEXcfi/bcOQLTZkxD7bcPy5N9Bvl1M4ftT1XHzqAXWPWbnaY5CZozNzYGYOXLN377rHqUT/jQdw4y3juPue+9j7M/vSt++GdY+kBi+VQ4Bpf53NTXdNZtCmfThg96346Z0PAnD9Hb9n4KYb1Dhha9Qeg3eDaVOnAtDW1sY5Z57BoYeNqHkizbdyj+6suvKKb97e8eMDeOSpl3h52ky22bIfANsN7s+UPy6fe62NfGlxCTv4wP359a/GM2P6dAZsuD5fP+VUZs+ew6WXXATAHnvuxWcPPLjmKTXfWu/tyTX/dRgA3bp25ZqxE7njnsc4+rUfcs5xe9OtWxdef30uI0+/uuZJqxeZWc2KI64GtgPWBP4MnJqZl3f0NVtsOTDHT7i3knlUjbU+/sW6R1AnvP7EtbS9NjUW9liVrybsV9W6JS15njOQBBgDSYUxkAQYA0mFMZAEGANJhTGQBBgDSYUxkAQYA0mFMZAEGANJhTGQBBgDSYUxkAQYA0mFMZAEGANJhTGQBBgDSYUxkAQYA0mFMZAEGANJhTGQBBgDSYUxkAQYA0mFMZAEGANJhTGQBBgDSYUxkAQYA0mFMZAEGANJhTGQBBgDSYUxkAQYA0mFMZAEGANJhTGQBBgDSYUxkAQYA0mFMZAEGANJhTGQBEBkZt0zvCkipgHP1T1HBdYEptc9hDplef0/2yAzey/sgaUqBsuriJiYmQPrnkPNezf+n3mYIAkwBpIKY9Aao+seQJ32rvs/85yBJMA9A0mFMZAEGINKRcSwiHgiIqZExIl1z6PFi4gxETE1Ih6ue5ZWMwYViYiuwIXALsAmwH4RsUm9U6kJVwDD6h6iDsagOoOBKZn5dGb+A/gRsGfNM2kxMvNu4C91z1EHY1CddYHnG+6/UJZJSyVjIAkwBlV6EfhAw/31yjJpqWQMqnMf0C8i+kZEd2Bf4KaaZ5IWyRhUJDPnAiOBccBjwLWZ+Ui9U2lxIuJq4DfAxhHxQkQcWvdMreLlyJIA9wwkFcZAEmAMJBXGQBJgDCQVxuBdKiK2i4hbyu09OnpXZUS8JyKOegfb+GZEfK3Z5Qs854qI2LsT2+rzbnyn4ZJkDJYz5d2SnZKZN2XmmR085T1Ap2OgZYsxWEaUn3yPR8RVEfFYRFwXESuXx56NiLMi4gFgn4jYOSJ+ExEPRMSPI2LV8rxhZR0PAJ9uWPdBEfHdcvt9EfGTiJhU/mwNnAlsGBEPRsQ55XnHRcR9ETE5Ik5rWNfJEfFkRPwa2LiJ7+uwsp5JEXH9/O+p2DEiJpb17V6e3zUizmnY9uH/7L+t2hmDZcvGwEWZ+SHgVd7+03pGZm4B/Bz4d2DHcn8i8JWI6AFcCnwK2BJYexHb+G9gfGZ+BNgCeAQ4EXgqMzfPzOMiYmegH+1v094c2DIito2ILWm/7HpzYFdgUBPf0w2ZOahs7zGg8Yq/PmUbuwHfK9/DocDMzBxU1n9YRPRtYjtajG51D6BOeT4zJ5TbVwJfBEaV+9eUvz9G+y9TmRARAN1pv7x2APBMZv4BICKuBEYsZBs7AAcCZOY8YGZE9FrgOTuXP78v91elPQ49gZ9k5mtlG828F+PDEXE67Yciq9J++fZ812ZmG/CHiHi6fA87A5s1nE9YvWz7ySa2pQ4Yg2XLgteON96fU/4O4I7M3K/xiRGx+RKcI4BvZ+YlC2zj2HewriuA4Zk5KSIOArZreGxh328Ax2RmYzSIiD7vYNtq4GHCsmX9iPh4ub0/8OuFPOe3wJCI2AggIlaJiP7A40CfiNiwPG+/hXwtwJ3AkeVru0bE6sAs2n/qzzcOOKThXMS6EbEWcDcwPCJWioietB+SLE5P4OWIWAE4YIHH9omILmXmDwJPlG0fWZ5PRPSPiFWa2I4WwxgsW54Ajo6Ix4BewMULPiEzpwEHAVdHxGTKIUJm/p32w4JbywnEqYvYxpeA7SPiIeB+YJPMnEH7YcfDEXFOZv4M+CHwm/K864CemfkA7Ycrk4CxtL+Ne3FOAX4HTKA9WI3+CNxb1nVE+R4uAx4FHigvJV6Ce7hLhO9aXEaU3eBbMvPDdc+i5ZN7BpIA9wwkFe4ZSAKMgaTCGEgCjIGkwhhIAuD/ALzayeuQBcFuAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from mlxtend.plotting import plot_confusion_matrix\n",
    "plot_confusion_matrix(confusion_matrix(y_test,y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "dbd9e1be",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.74      0.75      0.75        73\n",
      "           1       0.76      0.75      0.76        77\n",
      "\n",
      "    accuracy                           0.75       150\n",
      "   macro avg       0.75      0.75      0.75       150\n",
      "weighted avg       0.75      0.75      0.75       150\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "33f42442",
   "metadata": {},
   "source": [
    "#### Define different thresholds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "0cc666ca",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import precision_score, recall_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "05729312",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5/5 [==============================] - 0s 2ms/step\n",
      "0.3 -->> 0.7471264367816092 0.8441558441558441\n",
      "0.35 -->> 0.7529411764705882 0.8311688311688312\n",
      "0.4 -->> 0.7560975609756098 0.8051948051948052\n",
      "0.45 -->> 0.7468354430379747 0.7662337662337663\n",
      "0.5 -->> 0.7532467532467533 0.7532467532467533\n",
      "0.55 -->> 0.7532467532467533 0.7532467532467533\n",
      "0.6 -->> 0.7631578947368421 0.7532467532467533\n",
      "0.65 -->> 0.7464788732394366 0.6883116883116883\n",
      "0.7 -->> 0.7464788732394366 0.6883116883116883\n",
      "0.75 -->> 0.78125 0.6493506493506493\n"
     ]
    }
   ],
   "source": [
    "y_pred=model.predict(X_test)\n",
    "for i in [0.30, 0.35, 0.40, 0.45, 0.50, 0.55, 0.60, 0.65, 0.70, 0.75]:\n",
    "  classes = [1 if val>i else 0 for val in y_pred]\n",
    "  print(i, \"-->>\", precision_score(y_test, classes), recall_score(y_test, classes))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "6e797ce7",
   "metadata": {},
   "outputs": [],
   "source": [
    "classes = [1 if val>0.55 else 0 for val in y_pred]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "c1aef376",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(confusion_matrix(y_test, classes))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aad94bf2",
   "metadata": {},
   "source": [
    "#### 1D CNN model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "35f6beaa",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras import Sequential\n",
    "from tensorflow.keras.layers import Flatten, Dense, Conv1D, BatchNormalization, Dropout, SpatialDropout1D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "0a901e00",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(598, 300)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "9bc7c8df",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((598, 300, 1), (150, 300, 1))"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train=X_train.reshape(X_train.shape[0],X_train.shape[1],1) # ow it's 2D, first one is batch size \n",
    "X_test=X_test.reshape(X_test.shape[0],X_test.shape[1],1) # now it's 2D, first one is batch size\n",
    "X_train.shape, X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "9c923237",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(300, 1)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "052897c8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv1d_1 (Conv1D)           (None, 299, 32)           96        \n",
      "                                                                 \n",
      " batch_normalization_1 (Batc  (None, 299, 32)          128       \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 299, 32)           0         \n",
      "                                                                 \n",
      " flatten_1 (Flatten)         (None, 9568)              0         \n",
      "                                                                 \n",
      " dense_5 (Dense)             (None, 64)                612416    \n",
      "                                                                 \n",
      " dropout_1 (Dropout)         (None, 64)                0         \n",
      "                                                                 \n",
      " dense_6 (Dense)             (None, 1)                 65        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 612,705\n",
      "Trainable params: 612,641\n",
      "Non-trainable params: 64\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model=Sequential()\n",
    "model.add(Conv1D(32,2,activation=\"relu\",input_shape=X_train[0].shape))\n",
    "model.add(BatchNormalization()) # set data between 0-1\n",
    "model.add(Dropout(0.2))\n",
    "\n",
    "model.add(Flatten())\n",
    "model.add(Dense(64,activation=\"relu\"))\n",
    "model.add(Dropout(0.3))\n",
    "model.add(Dense(1,activation=\"sigmoid\"))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "ac16874f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "19/19 [==============================] - 2s 59ms/step - loss: 0.7558 - accuracy: 0.5084 - val_loss: 0.6931 - val_accuracy: 0.4933\n",
      "Epoch 2/100\n",
      "19/19 [==============================] - 1s 33ms/step - loss: 0.6894 - accuracy: 0.5870 - val_loss: 0.6916 - val_accuracy: 0.5067\n",
      "Epoch 3/100\n",
      "19/19 [==============================] - 1s 32ms/step - loss: 0.6804 - accuracy: 0.5936 - val_loss: 0.6900 - val_accuracy: 0.5600\n",
      "Epoch 4/100\n",
      "19/19 [==============================] - 1s 42ms/step - loss: 0.6542 - accuracy: 0.6271 - val_loss: 0.6885 - val_accuracy: 0.6133\n",
      "Epoch 5/100\n",
      "19/19 [==============================] - 1s 31ms/step - loss: 0.6318 - accuracy: 0.6355 - val_loss: 0.6868 - val_accuracy: 0.6467\n",
      "Epoch 6/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.6234 - accuracy: 0.6555 - val_loss: 0.6851 - val_accuracy: 0.6133\n",
      "Epoch 7/100\n",
      "19/19 [==============================] - 1s 38ms/step - loss: 0.6008 - accuracy: 0.6823 - val_loss: 0.6830 - val_accuracy: 0.6400\n",
      "Epoch 8/100\n",
      "19/19 [==============================] - 1s 42ms/step - loss: 0.5680 - accuracy: 0.7291 - val_loss: 0.6805 - val_accuracy: 0.6333\n",
      "Epoch 9/100\n",
      "19/19 [==============================] - 1s 33ms/step - loss: 0.5829 - accuracy: 0.7107 - val_loss: 0.6777 - val_accuracy: 0.6333\n",
      "Epoch 10/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.5833 - accuracy: 0.6890 - val_loss: 0.6746 - val_accuracy: 0.6467\n",
      "Epoch 11/100\n",
      "19/19 [==============================] - 1s 30ms/step - loss: 0.5511 - accuracy: 0.7291 - val_loss: 0.6713 - val_accuracy: 0.6533\n",
      "Epoch 12/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.5449 - accuracy: 0.7107 - val_loss: 0.6676 - val_accuracy: 0.6867\n",
      "Epoch 13/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.5209 - accuracy: 0.7642 - val_loss: 0.6632 - val_accuracy: 0.6933\n",
      "Epoch 14/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.5123 - accuracy: 0.7575 - val_loss: 0.6589 - val_accuracy: 0.7133\n",
      "Epoch 15/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.5126 - accuracy: 0.7575 - val_loss: 0.6537 - val_accuracy: 0.7200\n",
      "Epoch 16/100\n",
      "19/19 [==============================] - 1s 30ms/step - loss: 0.5240 - accuracy: 0.7375 - val_loss: 0.6485 - val_accuracy: 0.7000\n",
      "Epoch 17/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.4850 - accuracy: 0.7893 - val_loss: 0.6434 - val_accuracy: 0.7267\n",
      "Epoch 18/100\n",
      "19/19 [==============================] - 1s 33ms/step - loss: 0.5011 - accuracy: 0.7742 - val_loss: 0.6368 - val_accuracy: 0.7200\n",
      "Epoch 19/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.4884 - accuracy: 0.7876 - val_loss: 0.6305 - val_accuracy: 0.7200\n",
      "Epoch 20/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.4744 - accuracy: 0.7876 - val_loss: 0.6238 - val_accuracy: 0.7000\n",
      "Epoch 21/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.4834 - accuracy: 0.7759 - val_loss: 0.6172 - val_accuracy: 0.7267\n",
      "Epoch 22/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.4501 - accuracy: 0.8177 - val_loss: 0.6106 - val_accuracy: 0.7067\n",
      "Epoch 23/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.4681 - accuracy: 0.7793 - val_loss: 0.6028 - val_accuracy: 0.7400\n",
      "Epoch 24/100\n",
      "19/19 [==============================] - 1s 27ms/step - loss: 0.4597 - accuracy: 0.8010 - val_loss: 0.5956 - val_accuracy: 0.7333\n",
      "Epoch 25/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.4528 - accuracy: 0.7910 - val_loss: 0.5896 - val_accuracy: 0.7200\n",
      "Epoch 26/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.4412 - accuracy: 0.8010 - val_loss: 0.5829 - val_accuracy: 0.7200\n",
      "Epoch 27/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.4427 - accuracy: 0.7943 - val_loss: 0.5767 - val_accuracy: 0.7067\n",
      "Epoch 28/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.4230 - accuracy: 0.8144 - val_loss: 0.5711 - val_accuracy: 0.7267\n",
      "Epoch 29/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.4551 - accuracy: 0.7659 - val_loss: 0.5672 - val_accuracy: 0.7133\n",
      "Epoch 30/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.4288 - accuracy: 0.8110 - val_loss: 0.5628 - val_accuracy: 0.7333\n",
      "Epoch 31/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.4245 - accuracy: 0.8244 - val_loss: 0.5595 - val_accuracy: 0.7133\n",
      "Epoch 32/100\n",
      "19/19 [==============================] - 1s 30ms/step - loss: 0.4028 - accuracy: 0.8278 - val_loss: 0.5566 - val_accuracy: 0.7333\n",
      "Epoch 33/100\n",
      "19/19 [==============================] - 1s 31ms/step - loss: 0.3945 - accuracy: 0.8244 - val_loss: 0.5513 - val_accuracy: 0.7133\n",
      "Epoch 34/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.4099 - accuracy: 0.8244 - val_loss: 0.5482 - val_accuracy: 0.7133\n",
      "Epoch 35/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.4122 - accuracy: 0.8110 - val_loss: 0.5447 - val_accuracy: 0.7267\n",
      "Epoch 36/100\n",
      "19/19 [==============================] - 1s 30ms/step - loss: 0.4022 - accuracy: 0.8244 - val_loss: 0.5412 - val_accuracy: 0.7200\n",
      "Epoch 37/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.3983 - accuracy: 0.8177 - val_loss: 0.5389 - val_accuracy: 0.7267\n",
      "Epoch 38/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.3799 - accuracy: 0.8395 - val_loss: 0.5375 - val_accuracy: 0.7333\n",
      "Epoch 39/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.4030 - accuracy: 0.8278 - val_loss: 0.5383 - val_accuracy: 0.7400\n",
      "Epoch 40/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.3868 - accuracy: 0.8211 - val_loss: 0.5395 - val_accuracy: 0.7333\n",
      "Epoch 41/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.3796 - accuracy: 0.8478 - val_loss: 0.5360 - val_accuracy: 0.7333\n",
      "Epoch 42/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.3981 - accuracy: 0.8261 - val_loss: 0.5357 - val_accuracy: 0.7133\n",
      "Epoch 43/100\n",
      "19/19 [==============================] - 1s 31ms/step - loss: 0.3904 - accuracy: 0.8244 - val_loss: 0.5342 - val_accuracy: 0.7333\n",
      "Epoch 44/100\n",
      "19/19 [==============================] - 1s 32ms/step - loss: 0.3872 - accuracy: 0.8328 - val_loss: 0.5345 - val_accuracy: 0.7267\n",
      "Epoch 45/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.3917 - accuracy: 0.8311 - val_loss: 0.5350 - val_accuracy: 0.7200\n",
      "Epoch 46/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.3801 - accuracy: 0.8395 - val_loss: 0.5312 - val_accuracy: 0.7200\n",
      "Epoch 47/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.3615 - accuracy: 0.8512 - val_loss: 0.5312 - val_accuracy: 0.7200\n",
      "Epoch 48/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.3616 - accuracy: 0.8528 - val_loss: 0.5315 - val_accuracy: 0.7200\n",
      "Epoch 49/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.3714 - accuracy: 0.8495 - val_loss: 0.5327 - val_accuracy: 0.7067\n",
      "Epoch 50/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.3642 - accuracy: 0.8395 - val_loss: 0.5315 - val_accuracy: 0.7133\n",
      "Epoch 51/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.3646 - accuracy: 0.8395 - val_loss: 0.5296 - val_accuracy: 0.7133\n",
      "Epoch 52/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.3537 - accuracy: 0.8629 - val_loss: 0.5306 - val_accuracy: 0.7067\n",
      "Epoch 53/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.3549 - accuracy: 0.8462 - val_loss: 0.5287 - val_accuracy: 0.7267\n",
      "Epoch 54/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.3423 - accuracy: 0.8712 - val_loss: 0.5298 - val_accuracy: 0.7067\n",
      "Epoch 55/100\n",
      "19/19 [==============================] - 1s 28ms/step - loss: 0.3466 - accuracy: 0.8512 - val_loss: 0.5285 - val_accuracy: 0.7000\n",
      "Epoch 56/100\n",
      "19/19 [==============================] - 1s 31ms/step - loss: 0.3460 - accuracy: 0.8579 - val_loss: 0.5322 - val_accuracy: 0.7000\n",
      "Epoch 57/100\n",
      "19/19 [==============================] - 1s 30ms/step - loss: 0.3447 - accuracy: 0.8595 - val_loss: 0.5335 - val_accuracy: 0.7200\n",
      "Epoch 58/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 26ms/step - loss: 0.3562 - accuracy: 0.8528 - val_loss: 0.5338 - val_accuracy: 0.7067\n",
      "Epoch 59/100\n",
      "19/19 [==============================] - 0s 25ms/step - loss: 0.3529 - accuracy: 0.8462 - val_loss: 0.5325 - val_accuracy: 0.7133\n",
      "Epoch 60/100\n",
      "19/19 [==============================] - 1s 32ms/step - loss: 0.3428 - accuracy: 0.8512 - val_loss: 0.5319 - val_accuracy: 0.7133\n",
      "Epoch 61/100\n",
      "19/19 [==============================] - 1s 36ms/step - loss: 0.3270 - accuracy: 0.8629 - val_loss: 0.5319 - val_accuracy: 0.7000\n",
      "Epoch 62/100\n",
      "19/19 [==============================] - 1s 53ms/step - loss: 0.3136 - accuracy: 0.8746 - val_loss: 0.5335 - val_accuracy: 0.7200\n",
      "Epoch 63/100\n",
      "19/19 [==============================] - 1s 58ms/step - loss: 0.3447 - accuracy: 0.8645 - val_loss: 0.5322 - val_accuracy: 0.7200\n",
      "Epoch 64/100\n",
      "19/19 [==============================] - 1s 52ms/step - loss: 0.3331 - accuracy: 0.8662 - val_loss: 0.5321 - val_accuracy: 0.7200\n",
      "Epoch 65/100\n",
      "19/19 [==============================] - 1s 42ms/step - loss: 0.3402 - accuracy: 0.8495 - val_loss: 0.5312 - val_accuracy: 0.7067\n",
      "Epoch 66/100\n",
      "19/19 [==============================] - 1s 33ms/step - loss: 0.3146 - accuracy: 0.8846 - val_loss: 0.5291 - val_accuracy: 0.7200\n",
      "Epoch 67/100\n",
      "19/19 [==============================] - 1s 35ms/step - loss: 0.3296 - accuracy: 0.8763 - val_loss: 0.5263 - val_accuracy: 0.7133\n",
      "Epoch 68/100\n",
      "19/19 [==============================] - 1s 46ms/step - loss: 0.3144 - accuracy: 0.8813 - val_loss: 0.5267 - val_accuracy: 0.7200\n",
      "Epoch 69/100\n",
      "19/19 [==============================] - 1s 42ms/step - loss: 0.3312 - accuracy: 0.8729 - val_loss: 0.5256 - val_accuracy: 0.7133\n",
      "Epoch 70/100\n",
      "19/19 [==============================] - 1s 43ms/step - loss: 0.2957 - accuracy: 0.8829 - val_loss: 0.5292 - val_accuracy: 0.7267\n",
      "Epoch 71/100\n",
      "19/19 [==============================] - 1s 36ms/step - loss: 0.3162 - accuracy: 0.8729 - val_loss: 0.5286 - val_accuracy: 0.7133\n",
      "Epoch 72/100\n",
      "19/19 [==============================] - 1s 36ms/step - loss: 0.3125 - accuracy: 0.8679 - val_loss: 0.5304 - val_accuracy: 0.7200\n",
      "Epoch 73/100\n",
      "19/19 [==============================] - 1s 43ms/step - loss: 0.2932 - accuracy: 0.8880 - val_loss: 0.5302 - val_accuracy: 0.7200\n",
      "Epoch 74/100\n",
      "19/19 [==============================] - 1s 30ms/step - loss: 0.3135 - accuracy: 0.8746 - val_loss: 0.5277 - val_accuracy: 0.7200\n",
      "Epoch 75/100\n",
      "19/19 [==============================] - 1s 48ms/step - loss: 0.3231 - accuracy: 0.8813 - val_loss: 0.5282 - val_accuracy: 0.7267\n",
      "Epoch 76/100\n",
      "19/19 [==============================] - 1s 33ms/step - loss: 0.3263 - accuracy: 0.8629 - val_loss: 0.5289 - val_accuracy: 0.7200\n",
      "Epoch 77/100\n",
      "19/19 [==============================] - 1s 37ms/step - loss: 0.3198 - accuracy: 0.8712 - val_loss: 0.5285 - val_accuracy: 0.7133\n",
      "Epoch 78/100\n",
      "19/19 [==============================] - 1s 32ms/step - loss: 0.3077 - accuracy: 0.8796 - val_loss: 0.5279 - val_accuracy: 0.7133\n",
      "Epoch 79/100\n",
      "19/19 [==============================] - 0s 27ms/step - loss: 0.3098 - accuracy: 0.8846 - val_loss: 0.5312 - val_accuracy: 0.7133\n",
      "Epoch 80/100\n",
      "19/19 [==============================] - 1s 32ms/step - loss: 0.3004 - accuracy: 0.8846 - val_loss: 0.5299 - val_accuracy: 0.7333\n",
      "Epoch 81/100\n",
      "19/19 [==============================] - 1s 36ms/step - loss: 0.3015 - accuracy: 0.8829 - val_loss: 0.5314 - val_accuracy: 0.7267\n",
      "Epoch 82/100\n",
      "19/19 [==============================] - 1s 34ms/step - loss: 0.2855 - accuracy: 0.8863 - val_loss: 0.5326 - val_accuracy: 0.7067\n",
      "Epoch 83/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.3029 - accuracy: 0.8796 - val_loss: 0.5301 - val_accuracy: 0.7200\n",
      "Epoch 84/100\n",
      "19/19 [==============================] - 0s 26ms/step - loss: 0.2938 - accuracy: 0.8779 - val_loss: 0.5294 - val_accuracy: 0.7200\n",
      "Epoch 85/100\n",
      "19/19 [==============================] - 0s 25ms/step - loss: 0.2972 - accuracy: 0.8796 - val_loss: 0.5283 - val_accuracy: 0.7400\n",
      "Epoch 86/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.2953 - accuracy: 0.8829 - val_loss: 0.5290 - val_accuracy: 0.7200\n",
      "Epoch 87/100\n",
      "19/19 [==============================] - 1s 30ms/step - loss: 0.2928 - accuracy: 0.8779 - val_loss: 0.5272 - val_accuracy: 0.7333\n",
      "Epoch 88/100\n",
      "19/19 [==============================] - 1s 27ms/step - loss: 0.2967 - accuracy: 0.8846 - val_loss: 0.5281 - val_accuracy: 0.7267\n",
      "Epoch 89/100\n",
      "19/19 [==============================] - 0s 26ms/step - loss: 0.3024 - accuracy: 0.8880 - val_loss: 0.5310 - val_accuracy: 0.7400\n",
      "Epoch 90/100\n",
      "19/19 [==============================] - 0s 26ms/step - loss: 0.2977 - accuracy: 0.8746 - val_loss: 0.5284 - val_accuracy: 0.7267\n",
      "Epoch 91/100\n",
      "19/19 [==============================] - 0s 26ms/step - loss: 0.3000 - accuracy: 0.8763 - val_loss: 0.5299 - val_accuracy: 0.7333\n",
      "Epoch 92/100\n",
      "19/19 [==============================] - 1s 27ms/step - loss: 0.2856 - accuracy: 0.8963 - val_loss: 0.5270 - val_accuracy: 0.7400\n",
      "Epoch 93/100\n",
      "19/19 [==============================] - 0s 26ms/step - loss: 0.2799 - accuracy: 0.9047 - val_loss: 0.5295 - val_accuracy: 0.7333\n",
      "Epoch 94/100\n",
      "19/19 [==============================] - 1s 27ms/step - loss: 0.2946 - accuracy: 0.8746 - val_loss: 0.5283 - val_accuracy: 0.7400\n",
      "Epoch 95/100\n",
      "19/19 [==============================] - 0s 25ms/step - loss: 0.2902 - accuracy: 0.8880 - val_loss: 0.5292 - val_accuracy: 0.7333\n",
      "Epoch 96/100\n",
      "19/19 [==============================] - 0s 26ms/step - loss: 0.2642 - accuracy: 0.9080 - val_loss: 0.5294 - val_accuracy: 0.7333\n",
      "Epoch 97/100\n",
      "19/19 [==============================] - 1s 29ms/step - loss: 0.2766 - accuracy: 0.8997 - val_loss: 0.5283 - val_accuracy: 0.7400\n",
      "Epoch 98/100\n",
      "19/19 [==============================] - 0s 26ms/step - loss: 0.2883 - accuracy: 0.8913 - val_loss: 0.5256 - val_accuracy: 0.7467\n",
      "Epoch 99/100\n",
      "19/19 [==============================] - 1s 27ms/step - loss: 0.2886 - accuracy: 0.8880 - val_loss: 0.5252 - val_accuracy: 0.7400\n",
      "Epoch 100/100\n",
      "19/19 [==============================] - 0s 25ms/step - loss: 0.2787 - accuracy: 0.8863 - val_loss: 0.5301 - val_accuracy: 0.7400\n"
     ]
    }
   ],
   "source": [
    "epochs=100\n",
    "model.compile(optimizer=Adam(learning_rate=0.00001),loss=\"binary_crossentropy\",metrics=[\"accuracy\"])\n",
    "history=model.fit(X_train,y_train,epochs=epochs,validation_data=(X_test,y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "75c3f52e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['loss', 'accuracy', 'val_loss', 'val_accuracy'])"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history.history.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "ffcf352e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_history(history,epoch):\n",
    "    epoch_range=range(1,epoch+1)\n",
    "    plt.plot(epoch_range,history.history['accuracy'])\n",
    "    plt.plot(epoch_range,history.history['val_accuracy'])\n",
    "    plt.title(\"Model Accuracy\")\n",
    "    plt.ylabel('Accuracy')\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.legend(['Train', 'Val'], loc='upper right')\n",
    "    plt.show()\n",
    "    \n",
    "    # Plot training & validation loss values\n",
    "    plt.plot(epoch_range, history.history['loss'])\n",
    "    plt.plot(epoch_range, history.history['val_loss'])\n",
    "    plt.title('Model loss')\n",
    "    plt.ylabel('Loss')\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.legend(['Train', 'Val'], loc='upper left')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "a438425d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_history(history,epochs)\n",
    "# we have low data less than one thousand"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "638e03c8",
   "metadata": {},
   "source": [
    "#### Save, load, store, and test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "a3fb194e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import load_model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "1e7cdaca",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save(\"cnn_example_model\",save_format=\"h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "1b067051",
   "metadata": {},
   "outputs": [],
   "source": [
    "loading_model=load_model(\"cnn_example_model\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "65f5e1cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "x=\"It's a great movie. I'd like to watch it again\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "d64e6727",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'it is a great movie i would like to watch it again'"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_clean=clean_data(x)\n",
    "x_clean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "8e72907a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 300)"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vec=get_vec(x_clean).reshape(1,-1)\n",
    "vec.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "dadfd9d2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 300)"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vec=scaler.transform(vec)\n",
    "vec.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "17635cf6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 300, 1)"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vec=vec.reshape(1,300,1)\n",
    "vec.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "32aa8896",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 92ms/step\n"
     ]
    }
   ],
   "source": [
    "y_pred=loading_model.predict(vec)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "e68c8086",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred=np.where(y_pred > 0.6, 1,0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "bc6ad8f0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[1]]\n"
     ]
    }
   ],
   "source": [
    "print(y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "650e1f0e",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}