Text_nn.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "VyBoRfFt1OFE"
   },
   "source": [
    "# Product Categorization\n",
    "\n",
    "## Part 5: Multi-Class Text Classification of products by using Neural Networks\n",
    "\n",
    "The project concerns product categorization (multi-class text classification) based on their description. In this case I have used Neural Network and Convolutional Neural Network to get more accurate predictions. To calculated I have used python and Keras, TensorFlow frameworks.\n",
    "\n",
    "### Dataset\n",
    "\n",
    "The dataset comes from http://makeup-api.herokuapp.com/ and has been obtained by an API.\n",
    "\n",
    "The dataset contains the real descriptions about makeup products. Each description has been labeled with a specific product, therefore, this is a supervised text classification problem. \n",
    "\n",
    "Attributes:\n",
    "\n",
    "- product_type - category of makeup product.\n",
    "- description - description of makeup product."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import libriaries and data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "KfS8veRQ3_qg"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import LabelBinarizer, LabelEncoder\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Flatten, Activation, Dropout, GlobalMaxPool1D, Conv1D\n",
    "from keras.layers.embeddings import Embedding\n",
    "from keras.preprocessing import text, sequence\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "from keras.preprocessing.sequence import pad_sequences\n",
    "from keras import utils"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 204
    },
    "id": "FgbjCnBU80JM",
    "outputId": "909a18ea-d915-4486-87a5-ad02301f92fd"
   },
   "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>product_type</th>\n",
       "      <th>description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>lipstic</td>\n",
       "      <td>lippie pencil long wearing high intensity lip ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>lipstic</td>\n",
       "      <td>blotted lip sheer matte lipstick creates perfe...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>lipstic</td>\n",
       "      <td>lippie stix formula contains vitamin e mango a...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>foundation</td>\n",
       "      <td>developed selfie age buildable full coverage n...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>lipstic</td>\n",
       "      <td>product free lead heavy metal parabens phthala...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  product_type                                        description\n",
       "0      lipstic  lippie pencil long wearing high intensity lip ...\n",
       "1      lipstic  blotted lip sheer matte lipstick creates perfe...\n",
       "2      lipstic  lippie stix formula contains vitamin e mango a...\n",
       "3   foundation  developed selfie age buildable full coverage n...\n",
       "4      lipstic  product free lead heavy metal parabens phthala..."
      ]
     },
     "execution_count": 3,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('data\\products_clean.csv', header=0,index_col=0)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Shape of data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "id": "dzNQio4V80EN",
    "outputId": "f6ce7ddc-fdde-4767-fc0a-0c77f9474fc7"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(906, 2)"
      ]
     },
     "execution_count": 4,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "W9rKGKpSBiZh"
   },
   "source": [
    "## Data preparation\n",
    "\n",
    "Spliting the data into train and test sets:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "IkHqZeEH8z90"
   },
   "outputs": [],
   "source": [
    "sentences = df['description']\n",
    "y = df['product_type']\n",
    "\n",
    "sentences_train, sentences_test, train_y, test_y = train_test_split(sentences, y, test_size=0.25, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3QAGoEuMFzdb"
   },
   "source": [
    "#### Word Embeddings\n",
    "\n",
    "Word Embedding is a representation of text where words that have the same meaning have a similar representation. It shows words in a coordinate system where related words, based on a corpus of relationships, are placed closer together. In the deep learning frameworks such as  Keras, this part is generally handled by an embedding layer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "id": "gKMCfclv8z3D",
    "outputId": "95bb57c8-0b98-455d-928e-3d3f7a372a19"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "lippie stix formula contains vitamin e mango avocado shea butter added comfort moisture none lippie formula contain nasty ingredient like parabens sulfate\n",
      "[55, 358, 162, 810, 22, 121, 117, 500, 22, 198, 265, 4, 194, 14, 272, 321, 698, 218, 214, 596, 402, 597, 84, 93, 632, 288, 296, 237, 413]\n"
     ]
    }
   ],
   "source": [
    "tokenize = Tokenizer(num_words=1000)\n",
    "tokenize.fit_on_texts(sentences_train)\n",
    "\n",
    "X_train = tokenize.texts_to_sequences(sentences_train)\n",
    "X_test = tokenize.texts_to_sequences(sentences_test)\n",
    "\n",
    "vocab_size = len(tokenize.word_index) + 1  \n",
    "\n",
    "print(sentences_train[2])\n",
    "print(X_train[2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 119
    },
    "id": "GZKPGXLlCADP",
    "outputId": "e399dc65-8fa4-4d00-a647-6de88fadffac"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[148 278  22   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0   0]\n"
     ]
    }
   ],
   "source": [
    "maxlen = 100\n",
    "\n",
    "X_train = pad_sequences(X_train, padding='post', maxlen=maxlen)\n",
    "X_test = pad_sequences(X_test, padding='post', maxlen=maxlen)\n",
    "\n",
    "print(X_train[0, :])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "GhK21EV2I9VQ"
   },
   "outputs": [],
   "source": [
    "encoder = LabelEncoder()\n",
    "encoder.fit(train_y)\n",
    "y_train = encoder.transform(train_y)\n",
    "y_test = encoder.transform(test_y)\n",
    "\n",
    "num_classes = np.max(y_train) + 1\n",
    "y_train = utils.to_categorical(y_train, num_classes)\n",
    "y_test = utils.to_categorical(y_test, num_classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "QKAWaiZuHjm6"
   },
   "source": [
    "## Build model\n",
    "\n",
    "### Building the neural network model\n",
    "\n",
    "In this step I built and trained a several models to choose best of them. First of them is a simple neural network model.\n",
    "\n",
    "#### Model 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 289
    },
    "id": "Kcaj0Pp2CANk",
    "outputId": "caa48c5c-cee5-48cc-f0b8-f079c9130b6b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_8\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_8 (Embedding)      (None, 100, 16)           72304     \n",
      "_________________________________________________________________\n",
      "flatten_5 (Flatten)          (None, 1600)              0         \n",
      "_________________________________________________________________\n",
      "dense_16 (Dense)             (None, 16)                25616     \n",
      "_________________________________________________________________\n",
      "dense_17 (Dense)             (None, 5)                 85        \n",
      "=================================================================\n",
      "Total params: 98,005\n",
      "Trainable params: 98,005\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "embedding_dim = 16\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Embedding(input_dim=vocab_size, output_dim=embedding_dim, input_length=maxlen))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(16, activation='relu'))\n",
    "model.add(Dense(5, activation='softmax'))\n",
    "\n",
    "model.compile(optimizer='adam',\n",
    "              loss='categorical_crossentropy',\n",
    "              metrics=['accuracy'])\n",
    "model.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 533
    },
    "id": "yXEsYFuNYnJT",
    "outputId": "ae9cd484-699e-42f0-b9d0-5cb595a32700"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 10,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from keras.utils import plot_model\n",
    "plot_model(model, to_file='model_1.png', show_shapes=True, show_layer_names=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Training the model for 15 epoch:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 527
    },
    "id": "roe26IFLCAJk",
    "outputId": "7400fe18-5b23-4db0-c2e3-353beaa25fb8"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "22/22 [==============================] - 0s 9ms/step - loss: 1.5763 - accuracy: 0.3682 - val_loss: 1.5269 - val_accuracy: 0.4493\n",
      "Epoch 2/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 1.5031 - accuracy: 0.4330 - val_loss: 1.4760 - val_accuracy: 0.4714\n",
      "Epoch 3/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 1.4186 - accuracy: 0.4801 - val_loss: 1.3770 - val_accuracy: 0.4758\n",
      "Epoch 4/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 1.2379 - accuracy: 0.5950 - val_loss: 1.1745 - val_accuracy: 0.5374\n",
      "Epoch 5/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 0.9722 - accuracy: 0.7776 - val_loss: 0.9377 - val_accuracy: 0.7313\n",
      "Epoch 6/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 0.7018 - accuracy: 0.8645 - val_loss: 0.7600 - val_accuracy: 0.8062\n",
      "Epoch 7/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 0.4906 - accuracy: 0.9219 - val_loss: 0.6122 - val_accuracy: 0.8634\n",
      "Epoch 8/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 0.3475 - accuracy: 0.9485 - val_loss: 0.5245 - val_accuracy: 0.8678\n",
      "Epoch 9/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 0.2488 - accuracy: 0.9691 - val_loss: 0.4664 - val_accuracy: 0.8899\n",
      "Epoch 10/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 0.1884 - accuracy: 0.9764 - val_loss: 0.4194 - val_accuracy: 0.8855\n",
      "Epoch 11/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 0.1422 - accuracy: 0.9794 - val_loss: 0.3939 - val_accuracy: 0.8987\n",
      "Epoch 12/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 0.1096 - accuracy: 0.9867 - val_loss: 0.3656 - val_accuracy: 0.9119\n",
      "Epoch 13/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 0.0870 - accuracy: 0.9882 - val_loss: 0.3575 - val_accuracy: 0.9031\n",
      "Epoch 14/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 0.0698 - accuracy: 0.9971 - val_loss: 0.3471 - val_accuracy: 0.9075\n",
      "Epoch 15/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 0.0570 - accuracy: 0.9985 - val_loss: 0.3275 - val_accuracy: 0.9075\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train, y_train,\n",
    "                    batch_size=32,\n",
    "                    epochs=15,\n",
    "                    validation_data=(X_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predictions:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "id": "Vf2choBvCAIA",
    "outputId": "85c5b7d9-9c70-43d6-b883-831da8350195"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy: 0.9985\n",
      "Testing Accuracy:  0.9075\n"
     ]
    }
   ],
   "source": [
    "loss, accuracy = model.evaluate(X_train, y_train, verbose=False)\n",
    "print(\"Training Accuracy: {:.4f}\".format(accuracy))\n",
    "loss, accuracy = model.evaluate(X_test, y_test, verbose=False)\n",
    "print(\"Testing Accuracy:  {:.4f}\".format(accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Visualization of training:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 541
    },
    "id": "EG835YuIB__f",
    "outputId": "636b562b-338c-4ccc-d175-d3020183af8b"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_graphs(history, string):\n",
    "    plt.plot(history.history[string])\n",
    "    plt.plot(history.history['val_'+string])\n",
    "    plt.xlabel(\"Epochs\")\n",
    "    plt.ylabel(string)\n",
    "    plt.legend([string, 'val_'+string])\n",
    "    plt.show()\n",
    "\n",
    "plot_graphs(history, \"accuracy\")\n",
    "plot_graphs(history, \"loss\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "eBlERERxQx3D"
   },
   "source": [
    "At the end of the training one can see that there is a little bit of overfitting. \n",
    " \n",
    "#### Model 2\n",
    "In the second model I added dropout layers as a method of regularization."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 323
    },
    "id": "jfDsrM1uNkXF",
    "outputId": "68eb05b2-d20d-4175-be9f-dac2da555cd4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_7\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_7 (Embedding)      (None, 100, 16)           72304     \n",
      "_________________________________________________________________\n",
      "flatten_4 (Flatten)          (None, 1600)              0         \n",
      "_________________________________________________________________\n",
      "dense_14 (Dense)             (None, 16)                25616     \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 16)                0         \n",
      "_________________________________________________________________\n",
      "dense_15 (Dense)             (None, 5)                 85        \n",
      "=================================================================\n",
      "Total params: 98,005\n",
      "Trainable params: 98,005\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model2 = Sequential()\n",
    "model2.add(Embedding(input_dim=vocab_size, output_dim=embedding_dim, input_length=maxlen))\n",
    "model2.add(Flatten())\n",
    "model2.add(Dense(16, activation='relu'))\n",
    "model2.add(Dropout(0.5))\n",
    "model2.add(Dense(5, activation='softmax'))\n",
    "\n",
    "model2.compile(optimizer='adam',\n",
    "              loss='categorical_crossentropy',\n",
    "              metrics=['accuracy'])\n",
    "model2.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Training the model for 15 epoch:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 527
    },
    "id": "ZTgUrOswNkK5",
    "outputId": "5312da97-a1e3-4cd0-fd16-ca596b398f8f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "22/22 [==============================] - 0s 9ms/step - loss: 1.5486 - accuracy: 0.3402 - val_loss: 1.4551 - val_accuracy: 0.4537\n",
      "Epoch 2/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 1.5176 - accuracy: 0.3844 - val_loss: 1.4576 - val_accuracy: 0.4714\n",
      "Epoch 3/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 1.4749 - accuracy: 0.4183 - val_loss: 1.4237 - val_accuracy: 0.4714\n",
      "Epoch 4/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 1.4271 - accuracy: 0.4286 - val_loss: 1.3738 - val_accuracy: 0.4758\n",
      "Epoch 5/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 1.3298 - accuracy: 0.4507 - val_loss: 1.2525 - val_accuracy: 0.5066\n",
      "Epoch 6/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 1.2114 - accuracy: 0.5243 - val_loss: 1.0843 - val_accuracy: 0.6079\n",
      "Epoch 7/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 1.0125 - accuracy: 0.6200 - val_loss: 0.9304 - val_accuracy: 0.7753\n",
      "Epoch 8/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 0.8312 - accuracy: 0.7054 - val_loss: 0.7748 - val_accuracy: 0.7753\n",
      "Epoch 9/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 0.7146 - accuracy: 0.7673 - val_loss: 0.6667 - val_accuracy: 0.8018\n",
      "Epoch 10/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 0.6193 - accuracy: 0.7673 - val_loss: 0.5990 - val_accuracy: 0.8282\n",
      "Epoch 11/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 0.5493 - accuracy: 0.7938 - val_loss: 0.5476 - val_accuracy: 0.8282\n",
      "Epoch 12/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 0.4558 - accuracy: 0.8351 - val_loss: 0.5009 - val_accuracy: 0.8458\n",
      "Epoch 13/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 0.4299 - accuracy: 0.8365 - val_loss: 0.4665 - val_accuracy: 0.8458\n",
      "Epoch 14/15\n",
      "22/22 [==============================] - 0s 3ms/step - loss: 0.3762 - accuracy: 0.8748 - val_loss: 0.4486 - val_accuracy: 0.8590\n",
      "Epoch 15/15\n",
      "22/22 [==============================] - 0s 4ms/step - loss: 0.3693 - accuracy: 0.8439 - val_loss: 0.4249 - val_accuracy: 0.8678\n"
     ]
    }
   ],
   "source": [
    "history_two = model2.fit(X_train, y_train,\n",
    "                        batch_size=32,\n",
    "                        epochs=15,\n",
    "                        validation_data=(X_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predictions:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "id": "AZyc27y8HPFR",
    "outputId": "c26c18a9-1401-4669-d2dd-4da5cb4a0c2b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy: 0.9941\n",
      "Testing Accuracy:  0.8678\n"
     ]
    }
   ],
   "source": [
    "loss, accuracy = model2.evaluate(X_train, y_train, verbose=False)\n",
    "print(\"Training Accuracy: {:.4f}\".format(accuracy))\n",
    "loss, accuracy = model2.evaluate(X_test, y_test, verbose=False)\n",
    "print(\"Testing Accuracy:  {:.4f}\".format(accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Visualization of training:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 548
    },
    "id": "rh4E6tJFHPDj",
    "outputId": "689efc61-27b2-42b3-f8f3-fdb19a8a8a14"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEJCAYAAACZjSCSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deViU9frH8ffNjoICiqCAorkLrmhqaYtWuKTlvpWtZnXMltOpX/tyWk51OqfF8rSYS66plaZmZaaZKyqiuCsu4IaiuLJ/f388Y5ICDsowwNyv65oLZub7zNx4CZ95nu8mxhiUUkq5LjdnF6CUUsq5NAiUUsrFaRAopZSL0yBQSikXp0GglFIuToNAKaVcnMOCQETGicgREdlURJsbRSReRBJFZImjalFKKVU4cdQ8AhHpDJwGJhpjogp4PgBYDsQaY/aJSA1jzBGHFKOUUqpQHo56YWPMUhGJLKLJEGC2MWafrb1dIVC9enUTGVnUyyqllLrY2rVrjxpjggt6zmFBYIeGgKeI/Ab4Ax8YYyZe7qDIyEji4uIcXZtSSlUoIrK3sOecGQQeQBugC+ALrBCRlcaY7Rc3FJERwAiA2rVrl2qRSilV0Tlz1FAysNAYc8YYcxRYCrQoqKEx5jNjTIwxJiY4uMAzG6WUUlfImUHwPXC9iHiISCXgWmCLE+tRSimX5LBLQyIyFbgRqC4iycDLgCeAMWasMWaLiPwIJAB5wBfGmEKHmiqlXFt2djbJyclkZGQ4u5QyzcfHh/DwcDw9Pe0+xpGjhgbb0eZd4F1H1aCUqjiSk5Px9/cnMjISEXF2OWWSMYZjx46RnJxM3bp17T5OZxYrpcqFjIwMqlWrpiFQBBGhWrVqxT5r0iBQSpUbGgKXdyX/Ri4TBElHz/Dvn7bx+45UzmTmOLscpZQqM5w5j6BUbUpJ55PfdvHRrztxdxOa1apCu8gg2tYNom1kEEGVvZxdolKqjPPz8+P06dPOLqPEuUwQ3N6iFjc1rsG6vcdZsyeNVUlpTFy5ly+WJQHQoIYfbesG0S4yiHZ1g6gV4OvkipVSqnS4TBAA+Hl70LlhMJ0bWpPSMnNy2ZiczqqkNNbsSWNu/AGmrNoHQFiAL+3qWqHQNjKIa4Ir6/VJpRRgjc75xz/+wYIFCxARXnjhBQYOHMjBgwcZOHAgJ0+eJCcnh08//ZSOHTty//33ExcXh4hw33338cQTTzj7R/gLlwqCi3l7uBMTGURMZBAAuXmGLQdPsmaPFQy/70jl2/UpAFSr7EVb26WkdpFBNKnpj4e7y3SxKFWmvDo3kc0HTpboazatVYWXb29mV9vZs2cTHx/Phg0bOHr0KG3btqVz585MmTKF2267jeeff57c3FzOnj1LfHw8KSkpbNpkTZM6ceJEidZdElw6CC7m7iZEhVUlKqwq915XF2MMSUfPsDopjdW2cPgx8RBgnV20rhNIu8hAOlxTnVYRAbi56RmDUq5g2bJlDB48GHd3d0JCQrjhhhtYs2YNbdu25b777iM7O5s77riDli1bUq9ePXbv3s2oUaPo0aMHt956q7PLv4QGQRFEhHrBftQL9mNQO2uxu4Pp51htu5S0OimN935KBbYTWsWH2KhQukfXpE2dQNw1FJRyGHs/uZe2zp07s3TpUubNm8c999zDk08+yd13382GDRtYuHAhY8eOZcaMGYwbN87Zpf6FBkEx1azqS++WYfRuGQbA8TNZLNmeyvyNB5myeh/jl+8h2N+b2GahdIsOpV1kkF5CUqqC6dSpE//73/8YPnw4aWlpLF26lHfffZe9e/cSHh7Ogw8+SGZmJuvWraN79+54eXnRt29fGjVqxLBhw5xd/iU0CK5SYGUv7mgVxh2twjidmcPirUdYsOkg36zdz6SVe6lW2Ytbm4XSPTqU9vWq4amhoFS5d+edd7JixQpatGiBiPDOO+8QGhrKhAkTePfdd/H09MTPz4+JEyeSkpLCvffeS15eHgBvvfWWk6u/lMO2qnSUmJgYUx42pjmblcNv26wzhV+3HuFsVi6BlTy5tal1ptDxmup4eWgoKGWvLVu20KRJE2eXUS4U9G8lImuNMTEFtdczAgep5OVB9+iadI+uSUZ2Lku2p7Jg40HmbTzI9Lj9VPHx4JamofRoHsp19avj7eHu7JKVUi5Kg6AU+Hi6c1uzUG5rFkpGdi5/7DzKvI0H+WnzIWatS8bf24OuTUPoFhVK54bB+HhqKCilSo8GQSnz8XSnS5MQujQJISsnjz92HWXBxoP8tPkw365PobKXOzc3CaF7VCg3NqqBr5eGglLKsTQInMjLw42bGtXgpkY1eCM3jxW7jrFg00EWJh5m7oYD+Hl70CO6Jn1ah9E2MkjnKSilHEKDoIzwdHf7c/mL13vnsSopjW/Xp/BDwgGmx+0nPNCXPq3D6dMqjMjqlZ1drlKqAtEguJgxkJsNedm2rzlF3M+xvvoEQHBjcCuZUUAe7m5cV78619Wvzmu9m/FT4mFmrUvmo1938OGiHbSpE0jf1uH0aF6Tqr72b0enlFIFcZ0g2PkLLHyh6D/oudlgcq/s9StVh7qdoO4NUO8GCKwLJbBIXSUvjz/nKRxKz+C7+BRmrU3muW838srcRG5pEkLfNmF0ahCscxSUUlfEdYLAyx+q1wc3T3D3tH31yHffo5DHPcHNPd9zBbQ9dQh2L4GkJZD4rfV+VWtDvc5Q90ao2xn8Q676Rwit6sPIG67hoc712JRyklnrkpmz4QDzNh6kup8XvVuG0ad1GE1rVtGVUpVysqL2LtizZw89e/b8cyE6Z3OdIKh9rXVzlJZDrMtKR3dYgbD7N9gyF9Z/bT0f3MQ6U6jbGSKvB5+qV/xWIkJ0eFWiw6vyXPcmLNmeyux1yUxasZcvlyXRONSfvq3D6d2yFjWq+JTMz6eUqrBcJwhKgwgEN7Ru7R6EvFw4lHDhbGHtBFg1FsQNarW6cBkpoj14XtkfbC8PN25pGsItTUM4cTaLuQkHmbU2mTfmb+GtBVvo1CCYvm3CubVpiM5PUBXHgmfh0MaSfc3QaOj2dqFPP/vss0RERPDoo48C8Morr+Dh4cHixYs5fvw42dnZ/POf/6R3797FetuMjAwefvhh4uLi8PDw4P333+emm24iMTGRe++9l6ysLPLy8pg1axa1atViwIABJCcnk5uby4svvsjAgQOv6scGDQLHcnO3/uDXagXXPw45mZC85kIw/PEBLHsf3L2ts5W6N0C9G6FmS+uSUzEFVPLirvZ1uKt9HXalnubbdSl8uz6Fx6aux9/bgx7Na9KndThtIwP10pFSxTRw4EAef/zxP4NgxowZLFy4kMcee4wqVapw9OhR2rdvT69evYr1+zVmzBhEhI0bN7J161ZuvfVWtm/fztixYxk9ejRDhw4lKyuL3Nxc5s+fT61atZg3bx4A6enpJfKzaRCUJg9v67JQ5PXA85B5CvYuvxAMv75u3byrWG0adYdmd4C3f7Hf6ppgP/5+WyOevKUhK5OOMXtdCnM3HGDamv3c2jSEd/u1oGolHXGkyqkiPrk7SqtWrThy5AgHDhwgNTWVwMBAQkNDeeKJJ1i6dClubm6kpKRw+PBhQkND7X7dZcuWMWrUKAAaN25MnTp12L59Ox06dOCNN94gOTmZPn360KBBA6Kjo3nqqad45pln6NmzJ506dSqRn02HmTiTtz80vA1i34SH/4C/74R+46DZnXB4E8z5G7zXCL57FPausPogisnNTeh4TXXe69+CNS905dlujfl16xF6fPQ7G/aXvZ2SlCrL+vfvz8yZM5k+fToDBw5k8uTJpKamsnbtWuLj4wkJCSEjI6NE3mvIkCHMmTMHX19funfvzq+//krDhg1Zt24d0dHRvPDCC7z22msl8l4OCwIRGSciR0SkyG5xEWkrIjki0s9RtZQbfsEQ1Rd6fQijE+C+nyCqD2z+Dr6KhY9jYNl/rFFKV6CSlwcjb7iGGSM7YAz0G7ucr/5IorytQKuUswwcOJBp06Yxc+ZM+vfvT3p6OjVq1MDT05PFixezd+/eYr9mp06dmDx5MgDbt29n3759NGrUiN27d1OvXj0ee+wxevfuTUJCAgcOHKBSpUoMGzaMp59+mnXr1pXIz+XIM4LxQGxRDUTEHfgX8JMD6yifRKx+g94fw1PboPcYqBwMv7wC7zeFKYNg6zxr7kMxta4dyLzHrueGhsG8OnczI79eS/q54r+OUq6mWbNmnDp1irCwMGrWrMnQoUOJi4sjOjqaiRMn0rhx42K/5iOPPEJeXh7R0dEMHDiQ8ePH4+3tzYwZM4iKiqJly5Zs2rSJu+++m40bN9KuXTtatmzJq6++ygsvvFAiP5dD9yMQkUjgB2NMVCHPPw5kA21t7WZe7jXLy34EDnN0hzUkdcNUOH0YKteAFoOg1V3WaKViMMbw5bIk3l6wlZoBPnw8uDUtIgIcVLhSV0f3I7BfcfcjcFofgYiEAXcCn9rRdoSIxIlIXGpqquOLK8uqN4BbXoUnNsPgaRDRDlZ+AmPawhe3wLqJVie0HUSEBzrVY/pDHcjNNfQbu5zxeqlIKZfjzM7i/wLPGGPyLtfQGPOZMSbGGBMTHBxcCqWVA+4e0KgbDJoMT26BW16HjBMwZ1SxO5jb1Alk/uhO3NAwmFfmbubhr9fppSKlSsDGjRtp2bLlX27XXuvAia1XyGmXhkQkCTg/2LY6cBYYYYz5rqjXdPlLQ0UxxpqnsH4SbJoNWaehWn1oNQxaDAb/ooe06aUiVZZt2bKFxo0b6xyYyzDGsHXr1vJxacgYU9cYE2mMiQRmAo9cLgTUZYhYl4p6fWTrYP7k0g7mLT9AXsEnYXqpSJVlPj4+HDt2TP8/FsEYw7Fjx/DxKd5KBQ47IxCRqcCNWJ/2DwMvA54AxpixF7Udj3YWO87RnRD/NcRPhdOHoOkd0Odz8PAq9JATZ7N4asYGFm09QmyzUP7Vr7kuea2cKjs7m+Tk5BIbp19R+fj4EB4ejqfnX39fizojcOilIUfQILgKuTmw4mP45WW4pgsMnARehW9yY4zhi9+T+NeP1qWiMUNa0zxcLxUpVR6VyUtDygncPaw1j3p9BLsXw6Q+cK7w2cUiwoOdL1wq6vupXipSqiLSIHBFre+Gfl9BylqY0BNOFz0kt02dQOY91olODaxRRY9MXsfJDB1VpFRFoUHgqprdAUOmWf0HX8XCif1FNg+s7MUXd8fwXPfG/LT5MD0/XEZCsq5VpFRFoEHgyup3hbu/s84IxsVas5aL4OYmjOh8DTMeak9Obp5eKlKqgtAgcHW128M9P0BOhhUGBzdc9pA2dYL0UpFSFYgGgYKazeG+heDhA+N7WjOSL6OgS0W7Ugven1UpVbZpEChL9fpw34/gVwMm3Qk7frnsIfkvFZ3JzKH/2BW6x4FS5ZAGgbogIALu/dEKhamDIPFbuw5rUyeImQ93pJKXO4M/X8nvO1x8YUClyhkNAvVXfsEw/AcIj4GZ98HaCXYdVrd6ZWY93JHaQZW4b/wa5m444OBClVIlRYNAXco3AIbNhmtuhrmPwfKP7DospIoP0x/qQKuIQB6btp4Jy/c4tk6lVInQIFAF86oEg6Za6xL99AIset2uJa2r+noy8f52dGkcwstzEnn/5+06vFSpMk6DQBXOwwv6jbNmIv/+Hsx/utCVS/Pz8XRn7LDW9G8TzoeLdvD8d5vIzdMwUKqs8nB2AaqMc3OH2z8EnwBY/iFknrT2T3YveiVSD3c33unXnGp+3oxdsovjZ7L476CWeHu4l1LhSil7aRCoyxOBW16z+g4WvWZthdnvK/Ases1zEeHZbo2p7ufFP+dtIf2rNfzvrjb4++hy1kqVJXppSNlHBDo9Bd3fg23zYXI/u/dGfqBTPd4f0ILVSWkM/nwlqacyHVysUqo4NAhU8bR70NrUZu9ymNALzqbZdVif1uF8fncMO4+cpv/Y5exPO+vgQpVS9tIgUMXXfAAMmgyHE+Gr7nDyoF2H3dS4BpMfaM/xs9n0+XQ5Ww6edHChSil7aBCoK9OoGwybCen7YdxtkJZk12Ft6gTyzcgOuIsw4H8rWJ1k3xmFUspxNAjUlavbGYbPsUYSjYuF9GS7DmsY4s+sRzoS7O/NXV+u4ufNhx1cqFKqKBoE6uqEtYF75kPWGZgxHHKy7DsswJeZIzvSONSfhybFMWNN0RvjKKUcR4NAXb2QptD7Y0iJs2Yh2ymoshdTHmzPdfWr849ZCYxdsktnISvlBBoEqmQ0uwM6/A1W/w82zrT7sMreHnw5vC23t6jF2wu28sa8LeTpLGSlSpVOKFMlp+srkLIW5jwGIVFQo7Fdh3l5uPHBwJZUq+zFF8uSSDuTxb/6NcfTXT+nKFUa9DdNlRx3T2vGsVclmHGX3RPOwNrk5uXbm/L3Wxsye30KIybGcTYrx4HFKqXOc1gQiMg4ETkiIpsKeX6oiCSIyEYRWS4iLRxViypFVWpaYXBsp3VmUIxr/iLC325uwJt3RrNkeyrDvljFibP2dT4rpa6cI88IxgOxRTyfBNxgjIkGXgc+c2AtqjTV7QRdXoLE2bDqf8U+fMi1tflkaGs2pZxk2JeryMq5/IqnSqkr57AgMMYsBQqdLWSMWW6MOW67uxIId1QtygmuexwadYefnof9q4t9eGxUTT4c3IpNKScZu2SXAwpUSp1XVvoI7gcWOLsIVYJE4I5PoWq4Nb/gdPH3MY6NCuX2FrX46NcdbD9sf3+DUqp4nB4EInITVhA8U0SbESISJyJxqam6MXq54RsAAybBuTSYdT/k5Rb7JV65vSn+Pp78Y2aCbm6jlIM4NQhEpDnwBdDbGHOssHbGmM+MMTHGmJjg4ODSK1BdvZrNraWrk5bA4jeLfXg1P29evr0p8ftP8NUf9q1npJQqHqcFgYjUBmYDdxljtjurDlUKWt8Fre6ytrvcvrDYh/dqUYuuTWrw3k/b2HP0jAMKVMq1OXL46FRgBdBIRJJF5H4RGSkiI21NXgKqAZ+ISLyIxDmqFlUGdH8XQqNh9gg4vqdYh4oI/7wjGk83N56dnaAzj5UqYY4cNTTYGFPTGONpjAk3xnxpjBlrjBlre/4BY0ygMaal7RbjqFpUGeDpa/UXYGDG3ZCdUazDQ6v68HyPJqzcncbUNfscU6NSLsrpncXKhQTVhTv/Bwc3wI+Fjg0o1MC2EVxXvxpvzd/KgRPnHFCgUq5Jg0CVrkbd4PonYe14iJ9SrENFhLf7NCc3z/D8txt1pVKlSogGgSp9Nz0PkZ3ghyfg0MZiHRoRVImnb2vE4m2pfBef4qAClXItGgSq9Ll7QL9x4Bto9RdkpBfr8OEdI2ldO4BX524m9VSmg4pUynVoECjn8KsB/cfDiX3w3SPFWpzO3U14p19zzmbm8sqcRMfVqJSL0CBQzlO7PdzyGmz9AZZ/WKxD69fwZ3TXBszbeJAfNx1yUIFKuQYNAuVc7R+Bpr3hl1dhz7JiHTqicz2a1qzCi99vIv1stoMKVKri0yBQziUCvT62hpZ+cy+csv/Tvae7G+/0a07amSxen7fZgUUqVbFpECjn86liTTbLOm2FQa79n+6jwqoy8oZ6zFybzJLtuiChUldCg0CVDSFN4fYPYN9yWPRqsQ4ddXMDrgmuzHOzN3I6U7e3VKq4NAhU2dF8ALR9AJZ/BJvn2H2Yj6c77/RrwYH0c7zz41YHFqhUxaRBoMqW296EsDbw/aNwzP6dydrUCeSejpFMXLGX1UmFboynlCqABoEqWzy8of8EcPOA6XdB1lm7D336tkZEBPnyzKwEMrKLvwmOUq5Kg0CVPQER0PdzOLLZWobCzslmlbw8eLtPc5KOnuE/v+gWF0rZS4NAlU31u8KN/wcJ0+CPD+w+7Lr61RnUNoLPl+4mIfmEAwtUquLQIFBl1w3/gKi+8MsrsGWu3Yc916MJwf7e/GNmAlk5eY6rT6kKQoNAlV0i0HuM1Xk8ewQciLfrsCo+nrxxRzRbD53i09/s73BWylVpEKiyzdMXBk0B3yCYOghOHrTrsK5NQ+jVohYfL97BtkOnHFykUuWbBoEq+/xDYMh0yDxlhYGdI4levr0p/j6e/GNWArm6z7FShdIgUOVDaBT0/dLa5vLbhyDv8tf+q/l580qvZmzYf4Jxy5JKoUilyicNAlV+NIqF296ALXNg8T/tOuT25jXp2iSE937axp6jZxxcoFLlkwaBKl/aPwKth8Pv/4b4qZdtLiK8cWcUXh5uPDMrgTy9RKTUJTQIVPkiAj3+DXU7w9zHYO+Kyx4SUsWHF3o0YVVSGlNW7yuFIpUqXzQIVPnj7gkDJkJAbZg+FNIuf/1/QEwE19evztsLtnLgxLlSKFKp8kODQJVPvoEwZAbk5cKUgZCRXmRzEeGtPtHk5hme+3Yjphh7JCtV0TksCERknIgcEZFNhTwvIvKhiOwUkQQRae2oWlQFVe0aGPg1pO2ybWhT9F4EEUGVePq2Rvy2LZXZ61JKqUilyj5HnhGMB2KLeL4b0MB2GwF86sBaVEVVtxP0eB92LYKF/3fZ5sM7RtI2MpBX5iSyP83+lU2VqsgcFgTGmKVAUQvD9wYmGstKIEBEajqqHlWBtRkOHUfB6s9g1WdFNnV3E94f0BKAx6fHk5OraxEpZVcQiMhoEaliu5zzpYisE5Fbr/K9w4D9+e4n2x4r6P1HiEiciMSlpuq+tKoAXV+FRt3hx2dg5y9FNo0IqsQ/74xi7d7jfLx4ZykVqFTZZe8ZwX3GmJPArUAgcBfwtsOquogx5jNjTIwxJiY4OLi03laVJ27u0OdzqNHM6i84UvSWlb1bhnFnqzA+XLSDuD26o5lybfYGgdi+dgcmGWMS8z12pVKAiHz3w22PKXVlvP1gyDRrobopA+DM0SKbv9a7GWGBvoyeFs/JjOxSKlKpssfeIFgrIj9hBcFCEfEHrvbi6hzgbtvlpvZAujHGvqUllSpM1XAYNBVOH4ZpQyEns9Cm/j6efDCoFYdOZvDidwUOblPKJdgbBPcDzwJtjTFnAU/g3qIOEJGpwAqgkYgki8j9IjJSREbamswHdgM7gc+BR67kB1DqEuFt4I5PYf9KmPNYkVtdtq4dyOguDfg+/gDfrk8uxSKVKjs87GzXAYg3xpwRkWFAa6DI/QONMYMv87wBHrXz/ZUqnqg+cGyXtThdcEPo9FShTR+9qT6/70jlxe8SaVM7iNrVKpVioUo5n71nBJ8CZ0WkBfAUsAuY6LCqlCoJnf8O0QNg0Wuw+ftCm7m7Cf8Z2BIRGD19vQ4pVS7H3iDIsX2C7w18bIwZA/g7riylSoAI9PoIwtvB7IcgZV2hTcMDK/HGndGs33eCDxftKMUilXI+e4PglIj8H9aw0Xki4obVT6BU2ebpY2116RcMUwdDeuED03q1qEWf1mF8vHgnq5N0SKlyHfYGwUAgE2s+wSGsoZ7vOqwqpUqSXzAMng5ZZ2xbXRa+Qc1rvaMID6zEE9PjST+nQ0qVa7ArCGx//CcDVUWkJ5BhjNE+AlV+hDSF/l/B4U0we0ShW136eXvwwaCWHDqZwQvfbdJVSpVLsHeJiQHAaqA/MABYJSL9HFmYUiWuwS1w21uw9QdY9GqhzVrVDuSJrg2Yu+GArlKqXIK9w0efx5pDcARARIKBX4CZjipMKYe49iE4ug3++C8EN4aWBY9yfvjG+izdcZSXvt9ETGQgdapVLuVClSo99vYRuJ0PAZtjxThWqbJDBLq9c2Gry32rCmx2fkipu5swelo82TqkVFVg9v4x/1FEForIPSJyDzAPa2awUuWPuyf0n2AtRzFtCJwoeB/jsABf3uwTTfz+E3zwiw4pVRWXvZ3FTwOfAc1tt8+MMc84sjClHKpSkDWSKDfbGlaaebrAZj2b16Jfm3DG/LaTVbuPlXKRSpUOuy/vGGNmGWOetN2+dWRRSpWK4IYwYDwc2QKzHyx0JNErvZpRO8g2pPSsDilVFU+RQSAip0TkZAG3UyJysrSKVMphrrkZYt+GbfMLHUlkDSltxZFTmTz3nW58ryqeIoPAGONvjKlSwM3fGFOltIpUyqHaPQgx91kjieKnFtikZUQAT9zSkHkJB5m5VlcpVRWLjvxRys6RRCNvuIZr6wbx8pxE9hwtfHayUuWNBoFSYNdIovNDSj3d3Rg9bb0OKVUVhgaBUufZMZKoVoAvb/WJZkNyOv/5ebsTilSq5GkQKJVfcENrTaIjWwpdk6h7dE0GxITz6ZJdrNilQ0pV+adBoNTF6neB2Ldg2zz49bUCm7x8ezMiq1XmyRnxnDibVcoFKlWyNAiUKki7EdZIomX/KXAkUWXbKqWppzJ57lsdUqrKNw0CpQpix0ii5uEBPHVrI+ZvPMQ3cTqkVJVfGgRKFSb/SKLpQwscSfRQ53p0qFeNV+YmkqRDSlU5pUGgVFHOjyTKySpwJJGbm/D+wBZ/DinNytEhpar80SBQ6nIuM5KoZlVf/tU3moTkdB6ZvJZzWblOKlSpK6NBoJQ9LjOSKDaqJq/fEcWirUcY+sVKjp/RkUSq/HBoEIhIrIhsE5GdIvJsAc/XFpHFIrJeRBJEpLsj61HqqrQbAW3uLXQk0V3t6/Dp0NZsOnCSfmOXk3z8rBOKVKr4HBYEIuIOjAG6AU2BwSLS9KJmLwAzjDGtgEHAJ46qR6mrJgLd3y1yJFFsVE0m3deOI6cy6fvpcrYe0kV6VdnnyDOCdsBOY8xuY0wWMA3ofVEbA5xfxbQqcMCB9Sh19ewYSXRtvWp8M7IDgtB/7ApW6oY2qoxzZBCEAfvz3U+2PZbfK8AwEUnG2vpylAPrUapkXGYkEUDj0CrMeqQjIVV8uHvcahZsPOiEQpWyj7M7iwcD440x4UB3YJKIXFKTiIwQkTgRiUtNTS31IpW6xJ8jiTYXuiZRWIAvM0d2IKpWFR6Zso5JK/aUeplK2cORQZACROS7H257LL/7gRkAxpgVgA9Q/eIXMsZ8ZoyJMcbEBAcHO6hcpYqpfhfb7maFr0kUUMmLyQ+0p0vjEF78PpH3Fm7T5ShUmePIIFgDNBCRuiLihdUZPOeiNvuALgAi0gQrCMs4q8cAABW/SURBVPQjvyo/8o8k2jCtwCa+Xu6MHdaawe0i+HjxTp6ZlUCO7mWgyhAPR72wMSZHRP4GLATcgXHGmEQReQ2IM8bMAZ4CPheRJ7A6ju8x+nFJlSfnRxId2wlzRkGlatDglkuaebi78ead0QT7+/Dhoh0cPZ3FmCGt8fVyd0LRSv2VlLe/uzExMSYuLs7ZZSj1V2fTYGIvOJwIt74B7R+2QqIAX6/cy0vfb6JFRABfDm9LUGWvUi5WuSIRWWuMiSnoOWd3FitVMVQKgvsWQqPusPD/YO5oa1RRAYa1r8MnQ9uQaJt4tj9NJ54p59IgUKqkeFWGAZOg099h3QSYdKd1plCA2KhQJj9wLUdtE8+2HNSJZ8p5NAiUKklubtDlRejzBSSvgc9vgiNbC2zaNjKIb0Z2xE2EAWNX6LaXymk0CJRyhOb94d75kHUWvrwFdvxcYLNGof7MfqQjoVV9GD5uNfMSdOKZKn0aBEo5SngMjFgMgXVgygBY8QkUMDijVoAv34zsQIuIqvxt6jomLN9T+rUql6ZBoJQjVQ23qxM5oJIXk+6/lq5NQnh5TiLvLtyqE89UqdEgUMrR7OxE9vF059OhrRncrjZjFu/i6ZkJZOvEM1UKNAiUKg12diJbE8+ieKJrQ2auTWbExDjOZuU4oWDlSjQIlCpNdnQiiwijuzbgzTujWbI9lcGfr+LAiXNOKFa5Cg0CpUqbnZ3IQ66tzdhhbdh5+BTdPvidnxIPOaFY5Qo0CJRyhks6kR8rsBP51mahzHusExFBvoyYtJZX5iSSmZPrhIJVRaZBoJSz/KUTeaLViXzm0kllkdUrM+vhjtx/fV3GL99Dn0+Wszv10s1wlLpSGgRKOdOfncifW53IX9xcYCeyt4c7L/ZsypfDYzhw4hw9P1rG7HXJTihYVUQaBEqVBc0H2DUTuUuTEBaM7kx0WFWenLGBJ2fEcyZTRxWpq6NBoFRZYWcncmhVH6Y82J7Huzbgu/Up3P7RMjalpDuhYFVRaBAoVZbY2Yns7iY83rUhUx5sz5msHPp8spzxfyTpbGR1RTQIlCprLu5EntgLDm0qsGn7etVYMLoznRpU55W5mxkxaS3HzxS8D4JShdEgUKosyj8T+fBmGHs9zLwfju26pGlQZS++GB7DSz2b8tu2I3T/8HdWJxW8D4JSBdEgUKosa94fRsfD9U/AtvnwcVtrb+T0v44YEhHuu74usx++Di8PNwZ9toKPFu0gN08vFanL0z2LlSovTh2GZe9D3Djrfsz90OlJ8Kvx12YZ2bzw3Sa+jz9Ah3rV+O+gloRU8XFCwaosKWrPYg0CpcqbE/thyb8gfgp4eEP7h6HjKPAN/LOJMYaZa5N56ftEfL3c+feAFtzUqEYRL6oqOt28XqmKJCACen8Mj66GRt3g93/DBy1g6XuQac04FhH6x0Qwd9T11PD35t6v1vDGvM1k5eiy1upSekagVHl3aCP8+gZsXwCVg6HTU9DmXvC0LgdlZOfyxrwtTFq5lxbhVflocGtqV6vk5KJVadMzAqUqstBoGDIN7v8FajSBH5+Fj1rD2gmQm42Ppzuv3xHF2GGtSTp6hu4f/s7cDQecXbUqQzQIlKooItrC8Llw9/fgX9OajDamHWycCXl5xEbVZP7oTjQM8WPU1PU8OT2eg+m6z4FycBCISKyIbBORnSLybCFtBojIZhFJFJEpjqxHKZdQ70Z44BcYPA08fGHW/dY8hK3zCQ/wZfpDHRh1c31+SDjIje/+xtsLtpJ+NtvZVSsnclgfgYi4A9uBW4BkYA0w2BizOV+bBsAM4GZjzHERqWGMOVLU62ofgVLFkJcHibNh8ZuQtgvC2kCXl6DejexPO8t/ft7Ot/EpVPHx5NGbruHuDpH4eLo7u2rlAM7qI2gH7DTG7DbGZAHTgN4XtXkQGGOMOQ5wuRBQShWTmxtE97NGGPX6yJqLMLE3jO9JxMn1vD+gBfNGdaJlRABvzt/Kze/9xsy1yToRzcU4MgjCgP357ifbHsuvIdBQRP4QkZUiElvQC4nICBGJE5G41NRUB5WrVAXm7gGt74ZRayH2X5C6FcZ3hzHX0nTvJCYMqs+UB66lur83f/9mA90/+J1ftx7WRexchLM7iz2ABsCNwGDgcxEJuLiRMeYzY0yMMSYmODi4lEtUqgLx9IH2I2H0Buj1MXj7w8Ln4N+N6LjhWb7vkceYwa3IzMnlvvFxDPxsJev2HXd21crBHBkEKUBEvvvhtsfySwbmGGOyjTFJWH0KDRxYk1IKrBVOW98FDy6CkX9Am3tg+0/IhNvpsfR2FrXfwLvdarI79Qx9PlnOyElr2aXbY1ZYjuws9sD6w94FKwDWAEOMMYn52sRidSAPF5HqwHqgpTHm0o1bbbSzWCkHyToLm7+HteNh/0pw8ySnUQ/mut/CiwlBnMuBgW0jeLxLA2ro2kXljtPWGhKR7sB/AXdgnDHmDRF5DYgzxswREQH+DcQCucAbxphpRb2mBoFSpeDIVlg3wVrPKOMEuVXrsLhyN17c24LjboE8cH09RtxQjyo+ns6uVNlJF51TSl2Z7AzYMtc6S9i7DOPmwYZKHflPWgc2ebfikS6NGda+Nt4eOuS0rNMgUEpdvaM7LpwlnD1GqnsIEzM6sdQvlntiO9C7RRhubuLsKlUhNAiUUiUnJxO2zrNCYfdv5OLGr7mtWFalBzf1HMINjUOxrvqqskSDQCnlGGm7MWsnkhk3CZ/MoxwwQawP7MZ1fR4loHYzZ1en8tEgUEo5Vm422Vvmc/DXsYQdW4G7GE4HReHXdihE9QX/EGdX6PI0CJRSpWbrjh0smvkpnc79SnO3JIy4IfVuguYDoXEP8PZzdokuSYNAKVWqzmbl8PoPm1m9ZiUPBcTRx+MPPE4lg2claNzTCoV6N1pLX6hSoUGglHKK+RsP8uysBPLychnTKZvOGYuRxG8h44S1m1pUP2g+AGq1Au1gdigNAqWU0xw4cY7Hp8ezOimN3i1r8XrPBlTZ/xskTIftP0JuFlRrYJ0lNO8PgZHOLrlC0iBQSjlVbp5hzOKdfLBoB7UCfPhgUCta1w6Ec8dh8xxImAF7l1mNI9pbZwnN7oRKQc4tvALRIFBKlQlr96Yxelo8B9MzeKJrAx6+sT7u5yehndhnbauZMN1aJtvNExrcaoVCw1hr5VR1xTQIlFJlxsmMbJ7/dhNzNxzg2rpB/GdgS2oF+F5oYAwcSrDOEjZ+A6cPg3cVqHeD1cFc7yYIqqd9CsWkQaCUKlOMMcxal8JL32/C092Nf/WNJjaq5qUN83IhaQlsmg27FsPJZOvxqhG2YLgJ6nYGvxql+wOUQxoESqkyKenoGUZPW09CcjqD29XmpZ5N8fUqZAE7YyBtN+xeDLuXQNJSa/QRQI1mtrOFG6FOR52rUAANAqVUmZWVk8e/f97G/5bs5prgynw0uDVNa1W5/IF5uXBwA+z+zbrtWwm5meDmAeHtLgRDWGtw1+WyNQiUUmXesh1HeXJGPCfOZvNst8bce11k8Ravyz5nhcHu36zLSQfiAQNe/hB5/YVgCG7kkv0LGgRKqXLh2OlMnpmVwC9bjnBjo2De69+C6n7eV/ZiZ9Ngz+8XzhjSdluP+4Ve6Hiuc53V3+Dm7O3bHU+DQClVbhhjmLRyL/+ct4UqPp78e0ALbmgYfPUvfHyvdaaw+zerj+HsUetxd29rEltQXWs0UqDta1BdCKhdYS4raRAopcqdbYdOMWrqOrYfPs3919dlcLsI6lb3uzDv4Grk5cGRzbB/FRxPgjTb7XgSZJ+90E7crDOGgkIiMBK8Kl99LaVEg0ApVS5lZOfy5vwtTFyxF4BKXu40q1WFqLCqRIdVpXl41ZILB7BGJp0+fCEU0nbbQmK3df/c8b+29wu9KCRst2r1wadqydRUQjQIlFLl2q7U06zfd4JNKelsTEln84GTnMvOBUohHPI7d/yikNhz4ftTB//aNqA2hERDaBSENIOQKCssnNQfoUGglKpQcnLz2JV6ho0p6X+GQ+KBdDKy84BSDofzss7Cib1WKKRuhUOb4PAmOLYTjFUXXn5Qo2m+cIiGkKbg7e+4umw0CJRSFV6ZDAewAiJ1qxUK58Ph0CbITL/QJrCuLRxst9AoCKhTosNcNQiUUi7pcuEQUMmTfq3DGdq+DnWrl2LHrzGQvt8WDIlweKP1fdpuwPY32buK7azBdlkpNBpqNLniDmoNAqWUsskfDou3HmFh4iFy8gydGlTnrvZ1uLlxDTzcnTSvIPM0HNlinTX8eQaRCFmnrOevfRi6vX1FL+20IBCRWOADwB34whhT4E8gIn2BmUBbY0yRf+U1CJRSJenIyQymrdnPlFX7OHQyg1pVfRhybW0GtI2ghn8ZWPo6L8/qezi8yeqArtniil7GKUEgIu7AduAWIBlYAww2xmy+qJ0/MA/wAv6mQaCUcoac3Dx+2XKEyav28vuOo3i4CbFRodzVvg7t6gYVb7mLMqioIHDkztHtgJ3GmN22IqYBvYHNF7V7HfgX8LQDa1FKqSJ5uLsRGxVKbFQou1NPM3nVPr6J288PCQdpGOLHXe3rcEerMPx9KsZM4/wceSEsDNif736y7bE/iUhrIMIYM8+BdSilVLHUC/bjxZ5NWfVcV97p2xwvDzde/D6R9m8u4oXvNrL10Elnl1iiHHlGUCQRcQPeB+6xo+0IYARA7dq1HVuYUkrZ+Hq5M6BtBP1jwtmQnM6kFXuZEZfM1yv30S4yiGEd6hDbLBQvj/K9aJ0j+wg6AK8YY26z3f8/AGPMW7b7VYFdwGnbIaFAGtCrqH4C7SNQSjnT8TNZfLN2P1+v3Me+tLNU9/NiUNvaDL62NmH5t9wsY5zVWeyB1VncBUjB6iweYoxJLKT9b8DftbNYKVUe5OUZlu5I5euV+/h162EAbm4cwl0d6tCpfnXcHD1RrZic0llsjMkRkb8BC7GGj44zxiSKyGtAnDFmjqPeWymlHM3NTbixUQ1ubFSD5ONnmbp6H9NW7+eXLYfx8nDD39uDyt4e+Hl74Odj+2p7zN8n3/fn2+Vr4+fjgZ+XB5W93UtlToNOKFNKqRKSmZPLwsTDJB5I50xmDqczcjidmcvpzGxOZ+ZwJjOXUxk5nM7M/nN28+X4err/GR5Dr63NA53qXVFtzho+qpRSLsXbw51eLWrRq0Wty7bNyc2zgiEzmzO2sDiVkVPg91aY5Fz5bm2XoUGglFJO4OHuRtVKblSt5Px5CeV7zJNSSqmrpkGglFIuToNAKaVcnAaBUkq5OA0CpZRycRoESinl4jQIlFLKxWkQKKWUiyt3S0yISCqw9woPrw4cLcFyHK081VueaoXyVW95qhXKV73lqVa4unrrGGOCC3qi3AXB1RCRuMLW2iiLylO95alWKF/1lqdaoXzVW55qBcfVq5eGlFLKxWkQKKWUi3O1IPjM2QUUU3mqtzzVCuWr3vJUK5SvestTreCgel2qj0AppdSlXO2MQCml1EVcJghEJFZEtonIThF51tn1FEZEIkRksYhsFpFEERnt7JrsISLuIrJeRH5wdi1FEZEAEZkpIltFZIuIdHB2TUURkSds/w82ichUEfFxdk35icg4ETkiIpvyPRYkIj+LyA7b10Bn1nheIbW+a/u/kCAi34pIgDNrzK+gevM995SIGBGpXhLv5RJBICLuwBigG9AUGCwiTZ1bVaFygKeMMU2B9sCjZbjW/EYDW5xdhB0+AH40xjQGWlCGaxaRMOAxIMYYE4W19/cg51Z1ifFA7EWPPQssMsY0ABbZ7pcF47m01p+BKGNMc2A78H+lXVQRxnNpvYhIBHArsK+k3sglggBoB+w0xuw2xmQB04DeTq6pQMaYg8aYdbbvT2H9oQpzblVFE5FwoAfwhbNrKYqIVAU6A18CGGOyjDEnnFvVZXkAviLiAVQCDji5nr8wxiwF0i56uDcwwfb9BOCOUi2qEAXVaoz5yRiTY7u7Eggv9cIKUci/LcB/gH8AJdbB6ypBEAbsz3c/mTL+xxVARCKBVsAq51ZyWf/F+o9p327czlMXSAW+sl3G+kJEKju7qMIYY1KA97A++R0E0o0xPzm3KruEGGMO2r4/BIQ4s5hiuA9Y4OwiiiIivYEUY8yGknxdVwmCckdE/IBZwOPGmJPOrqcwItITOGKMWevsWuzgAbQGPjXGtALOUHYuW1zCdm29N1aA1QIqi8gw51ZVPMYalljmhyaKyPNYl2UnO7uWwohIJeA54KWSfm1XCYIUICLf/XDbY2WSiHhihcBkY8xsZ9dzGdcBvURkD9Ylt5tF5GvnllSoZCDZGHP+DGsmVjCUVV2BJGNMqjEmG5gNdHRyTfY4LCI1AWxfjzi5niKJyD1AT2CoKdvj6a/B+lCwwfb7Fg6sE5HQq31hVwmCNUADEakrIl5YHW5znFxTgUREsK5hbzHGvO/sei7HGPN/xphwY0wk1r/rr8aYMvmp1RhzCNgvIo1sD3UBNjuxpMvZB7QXkUq2/xddKMOd2/nMAYbbvh8OfO/EWookIrFYlzV7GWPOOrueohhjNhpjahhjIm2/b8lAa9v/66viEkFg6wz6G7AQ6xdphjEm0blVFeo64C6sT9bxtlt3ZxdVgYwCJotIAtASeNPJ9RTKduYyE1gHbMT6fS1TM2FFZCqwAmgkIskicj/wNnCLiOzAOqt525k1nldIrR8D/sDPtt+1sU4tMp9C6nXMe5XtMyGllFKO5hJnBEoppQqnQaCUUi5Og0AppVycBoFSSrk4DQKllHJxGgRK2YhIbr4hu/EluUqtiEQWtIqkUmWBh7MLUKoMOWeMaensIpQqbXpGoNRliMgeEXlHRDaKyGoRqW97PFJEfrWtZb9IRGrbHg+xrW2/wXY7vyyEu4h8bttf4CcR8bW1f8y2/0SCiExz0o+pXJgGgVIX+F50aWhgvufSjTHRWDNR/2t77CNggm0t+8nAh7bHPwSWGGNaYK1ldH4WewNgjDGmGXAC6Gt7/Fmgle11Rjrqh1OqMDqzWCkbETltjPEr4PE9wM3GmN22BQEPGWOqichRoKYxJtv2+EFjTHURSQXCjTGZ+V4jEvjZtlkLIvIM4GmM+aeI/AicBr4DvjPGnHbwj6rUX+gZgVL2MYV8XxyZ+b7P5UIfXQ+sHfRaA2tsm9AoVWo0CJSyz8B8X1fYvl/Oha0jhwK/275fBDwMf+7lXLWwFxURNyDCGLMYeAaoClxyVqKUI+knD6Uu8BWR+Hz3fzTGnB9CGmhbsTQTGGx7bBTWbmdPY+18dq/t8dHAZ7bVInOxQuEgBXMHvraFhQAfloPtM1UFo30ESl2GrY8gxhhz1Nm1KOUIemlIKaVcnJ4RKKWUi9MzAqWUcnEaBEop5eI0CJRSysVpECillIvTIFBKKRenQaCUUi7u/wHG2jit2TEKqAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_graphs(history_two, \"accuracy\")\n",
    "plot_graphs(history_two, \"loss\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "QehmKOWISfPM"
   },
   "source": [
    "#### Model 3\n",
    "Another way to work with embeddings is using a MaxPooling1D/AveragePooling1D or a GlobalMaxPooling1D/GlobalAveragePooling1D layer after the embedding. \n",
    "\n",
    "In this step I have used GlobalMaxPooling1D and add it to the sequential model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 289
    },
    "id": "odWssW7shNFH",
    "outputId": "dd4de5dd-14a4-4814-ed2e-f847031c4bab"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_1 (Embedding)      (None, 100, 50)           225950    \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d (Global (None, 50)                0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 16)                816       \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 5)                 85        \n",
      "=================================================================\n",
      "Total params: 226,851\n",
      "Trainable params: 226,851\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "embedding_dim = 50\n",
    "\n",
    "model3 = Sequential()\n",
    "model3.add(Embedding(input_dim=vocab_size, output_dim=embedding_dim, input_length=maxlen))\n",
    "model3.add(GlobalMaxPool1D())\n",
    "model3.add(Dense(16, activation='relu'))\n",
    "model3.add(Dense(5, activation='softmax'))\n",
    "\n",
    "model3.compile(optimizer='adam',\n",
    "               loss='categorical_crossentropy',\n",
    "               metrics=['accuracy'])\n",
    "\n",
    "model3.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Training for 15 epoch:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 527
    },
    "id": "4odNmbvthM6z",
    "outputId": "863c26a8-1cb5-4f17-8589-66c3062d7432"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "22/22 [==============================] - 0s 12ms/step - loss: 1.5949 - accuracy: 0.2091 - val_loss: 1.5811 - val_accuracy: 0.1806\n",
      "Epoch 2/15\n",
      "22/22 [==============================] - 0s 7ms/step - loss: 1.5638 - accuracy: 0.3726 - val_loss: 1.5460 - val_accuracy: 0.4890\n",
      "Epoch 3/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 1.5192 - accuracy: 0.5700 - val_loss: 1.4922 - val_accuracy: 0.6167\n",
      "Epoch 4/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 1.4554 - accuracy: 0.6009 - val_loss: 1.4153 - val_accuracy: 0.5991\n",
      "Epoch 5/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 1.3650 - accuracy: 0.6082 - val_loss: 1.3129 - val_accuracy: 0.6344\n",
      "Epoch 6/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 1.2474 - accuracy: 0.6627 - val_loss: 1.1749 - val_accuracy: 0.7533\n",
      "Epoch 7/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 1.0934 - accuracy: 0.7644 - val_loss: 1.0047 - val_accuracy: 0.8106\n",
      "Epoch 8/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.9255 - accuracy: 0.8189 - val_loss: 0.8532 - val_accuracy: 0.8458\n",
      "Epoch 9/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.7676 - accuracy: 0.8439 - val_loss: 0.7199 - val_accuracy: 0.8634\n",
      "Epoch 10/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.6398 - accuracy: 0.8748 - val_loss: 0.6217 - val_accuracy: 0.8634\n",
      "Epoch 11/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.5389 - accuracy: 0.8910 - val_loss: 0.5474 - val_accuracy: 0.8678\n",
      "Epoch 12/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.4588 - accuracy: 0.8969 - val_loss: 0.4902 - val_accuracy: 0.8811\n",
      "Epoch 13/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.3931 - accuracy: 0.9411 - val_loss: 0.4432 - val_accuracy: 0.8987\n",
      "Epoch 14/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.3358 - accuracy: 0.9573 - val_loss: 0.4039 - val_accuracy: 0.8987\n",
      "Epoch 15/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.2889 - accuracy: 0.9632 - val_loss: 0.3737 - val_accuracy: 0.9163\n"
     ]
    }
   ],
   "source": [
    "history_3 = model3.fit(X_train, y_train,\n",
    "                       batch_size=32,\n",
    "                       epochs=15,\n",
    "                       validation_data=(X_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predictions:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "id": "k5rwoZO6h5UW",
    "outputId": "abeb7f17-2fec-46bb-ded1-8d7cb67dca34"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy: 0.9735\n",
      "Testing Accuracy:  0.9163\n"
     ]
    }
   ],
   "source": [
    "loss, accuracy = model3.evaluate(X_train, y_train, verbose=False)\n",
    "print(\"Training Accuracy: {:.4f}\".format(accuracy))\n",
    "loss, accuracy = model3.evaluate(X_test, y_test, verbose=False)\n",
    "print(\"Testing Accuracy:  {:.4f}\".format(accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Visualization of training:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 545
    },
    "id": "gZRVXaW1h9Jc",
    "outputId": "3ff62d04-8e44-4a08-b558-96a3318a03c6"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_graphs(history_3, \"accuracy\")\n",
    "plot_graphs(history_3, \"loss\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Model 4: Final Neural Network model\n",
    "\n",
    "The final model consist of an Embedding layer with GlobalMaxPool1D layer, hidden layer using the ReLU activation function. The output layer using the softmax activation function. I have also used dropout layers as a method of regularization, Adam optimizer and categorical cross entropy loss as the loss function."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 357
    },
    "id": "6TEBDh_LSdfx",
    "outputId": "674ec372-d0fa-45c6-9c77-f30251d3c9c1"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_2 (Embedding)      (None, 100, 50)           225950    \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d_1 (Glob (None, 50)                0         \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 50)                0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 16)                816       \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 16)                0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 5)                 85        \n",
      "=================================================================\n",
      "Total params: 226,851\n",
      "Trainable params: 226,851\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "embedding_dim = 50\n",
    "\n",
    "model4 = Sequential()\n",
    "model4.add(Embedding(input_dim=vocab_size, output_dim=embedding_dim, input_length=maxlen))\n",
    "model4.add(GlobalMaxPool1D())\n",
    "model4.add(Dropout(0.2))\n",
    "model4.add(Dense(16, activation='relu'))\n",
    "model4.add(Dropout(0.2))\n",
    "model4.add(Dense(5, activation='softmax'))\n",
    "\n",
    "model4.compile(optimizer='adam',\n",
    "               loss='categorical_crossentropy',\n",
    "               metrics=['accuracy'])\n",
    "\n",
    "model4.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Training model for 15 epoch:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 527
    },
    "id": "J4qeOHeISdan",
    "outputId": "9bb64068-bc15-471e-9312-6b16aade7425"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "22/22 [==============================] - 0s 12ms/step - loss: 1.6062 - accuracy: 0.2224 - val_loss: 1.5875 - val_accuracy: 0.5022\n",
      "Epoch 2/15\n",
      "22/22 [==============================] - 0s 7ms/step - loss: 1.5768 - accuracy: 0.3711 - val_loss: 1.5546 - val_accuracy: 0.5066\n",
      "Epoch 3/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 1.5396 - accuracy: 0.4315 - val_loss: 1.5030 - val_accuracy: 0.4670\n",
      "Epoch 4/15\n",
      "22/22 [==============================] - 0s 7ms/step - loss: 1.4852 - accuracy: 0.4153 - val_loss: 1.4280 - val_accuracy: 0.4670\n",
      "Epoch 5/15\n",
      "22/22 [==============================] - 0s 7ms/step - loss: 1.4075 - accuracy: 0.4271 - val_loss: 1.3391 - val_accuracy: 0.4670\n",
      "Epoch 6/15\n",
      "22/22 [==============================] - 0s 7ms/step - loss: 1.3191 - accuracy: 0.4433 - val_loss: 1.2436 - val_accuracy: 0.5066\n",
      "Epoch 7/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 1.2069 - accuracy: 0.5464 - val_loss: 1.1282 - val_accuracy: 0.6300\n",
      "Epoch 8/15\n",
      "22/22 [==============================] - 0s 7ms/step - loss: 1.0861 - accuracy: 0.6510 - val_loss: 0.9843 - val_accuracy: 0.6872\n",
      "Epoch 9/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.9219 - accuracy: 0.7261 - val_loss: 0.8399 - val_accuracy: 0.7885\n",
      "Epoch 10/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.7975 - accuracy: 0.7791 - val_loss: 0.7157 - val_accuracy: 0.8282\n",
      "Epoch 11/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.6659 - accuracy: 0.8262 - val_loss: 0.6150 - val_accuracy: 0.8590\n",
      "Epoch 12/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.5937 - accuracy: 0.8292 - val_loss: 0.5385 - val_accuracy: 0.8811\n",
      "Epoch 13/15\n",
      "22/22 [==============================] - 0s 7ms/step - loss: 0.5148 - accuracy: 0.8542 - val_loss: 0.4784 - val_accuracy: 0.9031\n",
      "Epoch 14/15\n",
      "22/22 [==============================] - 0s 7ms/step - loss: 0.4588 - accuracy: 0.8792 - val_loss: 0.4361 - val_accuracy: 0.9119\n",
      "Epoch 15/15\n",
      "22/22 [==============================] - 0s 6ms/step - loss: 0.4188 - accuracy: 0.8763 - val_loss: 0.3942 - val_accuracy: 0.9339\n"
     ]
    }
   ],
   "source": [
    "history_4 = model4.fit(X_train, y_train,\n",
    "                       batch_size=32,\n",
    "                       epochs=15,\n",
    "                       validation_data=(X_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predictions:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "id": "tf-gE9K5qq0-",
    "outputId": "2e406001-83a5-4cdd-8cd7-c852ee3f6ef3"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy: 0.9543\n",
      "Testing Accuracy:  0.9339\n"
     ]
    }
   ],
   "source": [
    "loss, accuracy = model4.evaluate(X_train, y_train, verbose=False)\n",
    "print(\"Training Accuracy: {:.4f}\".format(accuracy))\n",
    "loss, accuracy = model4.evaluate(X_test, y_test, verbose=False)\n",
    "print(\"Testing Accuracy:  {:.4f}\".format(accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Visualization of training:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 541
    },
    "id": "skqAozltquzR",
    "outputId": "027bc033-8e61-4e2e-bd61-b6c0a7236be5"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_graphs(history_4, \"accuracy\")\n",
    "plot_graphs(history_4, \"loss\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "bAT2Y3SjaCjd"
   },
   "source": [
    "From above calculation one can see that I have obtained a pretty good model.\n",
    "\n",
    "### Convolutional Neural Network (CNN)\n",
    "\n",
    "I decided to build a CNN model to check that I will achive a better result that previous calculation. The model contains a simple CNN architecture: Embedding layer with single layer Convolutional Neural Network (with RELU activation), and the GlobalMaxPool1D layer as well. The hidden layer using the ReLU activation function and an output layer using the softmax activation function. I also have used the Adam optimizer and categorical cross entropy loss as the loss function."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 323
    },
    "id": "GF0GK9lrquvH",
    "outputId": "b0c5a27b-d038-45b8-a4ef-3c7f0dd978a0"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_20\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_20 (Embedding)     (None, 100, 50)           225950    \n",
      "_________________________________________________________________\n",
      "conv1d_11 (Conv1D)           (None, 96, 128)           32128     \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d_13 (Glo (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_40 (Dense)             (None, 10)                1290      \n",
      "_________________________________________________________________\n",
      "dense_41 (Dense)             (None, 5)                 55        \n",
      "=================================================================\n",
      "Total params: 259,423\n",
      "Trainable params: 259,423\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "embedding_dim = 50\n",
    "\n",
    "model5 = Sequential()\n",
    "model5.add(Embedding(input_dim=vocab_size, output_dim=embedding_dim, input_length=maxlen))\n",
    "model5.add(Conv1D(128, 5, activation='relu'))\n",
    "model5.add(GlobalMaxPool1D())\n",
    "model5.add(Dense(10, activation='relu'))\n",
    "model5.add(Dense(5, activation='softmax'))\n",
    "\n",
    "model5.compile(optimizer='adam',\n",
    "               loss='categorical_crossentropy',\n",
    "               metrics=['accuracy'])\n",
    "\n",
    "model5.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Training NN model for 5 epoch:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 187
    },
    "id": "QUCJJZbQqusm",
    "outputId": "0169a515-3afc-4ac9-b672-2f3b6cd0ce6b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "68/68 [==============================] - 1s 14ms/step - loss: 1.5243 - accuracy: 0.3608 - val_loss: 1.4174 - val_accuracy: 0.4714\n",
      "Epoch 2/5\n",
      "68/68 [==============================] - 1s 11ms/step - loss: 1.2113 - accuracy: 0.5611 - val_loss: 0.8595 - val_accuracy: 0.7797\n",
      "Epoch 3/5\n",
      "68/68 [==============================] - 1s 11ms/step - loss: 0.6216 - accuracy: 0.8645 - val_loss: 0.4337 - val_accuracy: 0.9075\n",
      "Epoch 4/5\n",
      "68/68 [==============================] - 1s 12ms/step - loss: 0.2523 - accuracy: 0.9602 - val_loss: 0.2886 - val_accuracy: 0.9163\n",
      "Epoch 5/5\n",
      "68/68 [==============================] - 1s 11ms/step - loss: 0.0983 - accuracy: 0.9823 - val_loss: 0.2321 - val_accuracy: 0.9295\n"
     ]
    }
   ],
   "source": [
    "history_5 = model5.fit(X_train, y_train,\n",
    "                       batch_size=10,\n",
    "                       epochs=5,\n",
    "                       validation_data=(X_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predictions:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 51
    },
    "id": "ueVfCA9mquqt",
    "outputId": "dc8cf033-7204-4040-cebf-d4cdad03dea9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy: 0.9956\n",
      "Testing Accuracy:  0.9295\n"
     ]
    }
   ],
   "source": [
    "loss, accuracy = model5.evaluate(X_train, y_train, verbose=False)\n",
    "print(\"Training Accuracy: {:.4f}\".format(accuracy))\n",
    "loss, accuracy = model5.evaluate(X_test, y_test, verbose=False)\n",
    "print(\"Testing Accuracy:  {:.4f}\".format(accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Visualization of training:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 541
    },
    "id": "umUCVe_KaS0C",
    "outputId": "ef41571a-c42b-4a42-95d8-59f01e5bc781"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_graphs(history_5, \"accuracy\")\n",
    "plot_graphs(history_5, \"loss\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "LYcp0sGTdB-t"
   },
   "source": [
    "The model was received a similar results to simple NN model. The accuracy was better than previous model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Summary\n",
    "\n",
    "The project was aimed product categorization based on their descriptionn. In this approach I have used Neural Network and Convolutional Neural Network to resolve the text classification problem. I build and trained 4 models by using simple Neural Network and the best model achieved an accuracy on the test set equal to 95 %. Additionaly I built the CNN model and it was received a similar results to simple NN models. "
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "Text_nn.ipynb",
   "provenance": []
  },
  "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": "3.12.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}