Titanic_Perceptron_KNN_MLP_Adaline.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Titanic-Perceptron-KNN-MLP-Adaline",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "Eld4miMCOcXr"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import tensorflow as tf\n",
        "from numpy.linalg import inv\n",
        "from tensorflow.keras.layers import Dense"
      ],
      "execution_count": 108,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zUxwz7c3yEaz",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "db1b8511-216e-41be-a1b2-07492d7871c4"
      },
      "source": [
        "train_data = pd.read_csv('/content/drive/MyDrive/csv_files/train.csv')\n",
        "test_data = pd.read_csv('/content/drive/MyDrive/csv_files/test.csv')\n",
        "gender_submission = pd.read_csv('/content/drive/MyDrive/csv_files/gender_submission.csv')\n",
        "train_data.head()"
      ],
      "execution_count": 109,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>PassengerId</th>\n",
              "      <th>Survived</th>\n",
              "      <th>Pclass</th>\n",
              "      <th>Name</th>\n",
              "      <th>Sex</th>\n",
              "      <th>Age</th>\n",
              "      <th>SibSp</th>\n",
              "      <th>Parch</th>\n",
              "      <th>Ticket</th>\n",
              "      <th>Fare</th>\n",
              "      <th>Cabin</th>\n",
              "      <th>Embarked</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>Braund, Mr. Owen Harris</td>\n",
              "      <td>male</td>\n",
              "      <td>22.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>A/5 21171</td>\n",
              "      <td>7.2500</td>\n",
              "      <td>NaN</td>\n",
              "      <td>S</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>Cumings, Mrs. John Bradley (Florence Briggs Th...</td>\n",
              "      <td>female</td>\n",
              "      <td>38.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>PC 17599</td>\n",
              "      <td>71.2833</td>\n",
              "      <td>C85</td>\n",
              "      <td>C</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>Heikkinen, Miss. Laina</td>\n",
              "      <td>female</td>\n",
              "      <td>26.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>STON/O2. 3101282</td>\n",
              "      <td>7.9250</td>\n",
              "      <td>NaN</td>\n",
              "      <td>S</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>Futrelle, Mrs. Jacques Heath (Lily May Peel)</td>\n",
              "      <td>female</td>\n",
              "      <td>35.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>113803</td>\n",
              "      <td>53.1000</td>\n",
              "      <td>C123</td>\n",
              "      <td>S</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>Allen, Mr. William Henry</td>\n",
              "      <td>male</td>\n",
              "      <td>35.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>373450</td>\n",
              "      <td>8.0500</td>\n",
              "      <td>NaN</td>\n",
              "      <td>S</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   PassengerId  Survived  Pclass  ...     Fare Cabin  Embarked\n",
              "0            1         0       3  ...   7.2500   NaN         S\n",
              "1            2         1       1  ...  71.2833   C85         C\n",
              "2            3         1       3  ...   7.9250   NaN         S\n",
              "3            4         1       1  ...  53.1000  C123         S\n",
              "4            5         0       3  ...   8.0500   NaN         S\n",
              "\n",
              "[5 rows x 12 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 109
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        },
        "id": "yePZ1p-PF9go",
        "outputId": "df7fae3a-bb36-430c-bd27-0bee3a5ceec0"
      },
      "source": [
        "train_data = train_data.replace([\"female\", \"male\"], [0, 1])\n",
        "train_data = train_data.replace([\"S\", \"C\", \"Q\"], [0, 1, 2])\n",
        "test_data = test_data.replace([\"female\", \"male\"], [0, 1])\n",
        "test_data = test_data.replace([\"S\", \"C\", \"Q\"], [0, 1, 2])\n",
        "train_data"
      ],
      "execution_count": 110,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>PassengerId</th>\n",
              "      <th>Survived</th>\n",
              "      <th>Pclass</th>\n",
              "      <th>Name</th>\n",
              "      <th>Sex</th>\n",
              "      <th>Age</th>\n",
              "      <th>SibSp</th>\n",
              "      <th>Parch</th>\n",
              "      <th>Ticket</th>\n",
              "      <th>Fare</th>\n",
              "      <th>Cabin</th>\n",
              "      <th>Embarked</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>Braund, Mr. Owen Harris</td>\n",
              "      <td>1</td>\n",
              "      <td>22.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>A/5 21171</td>\n",
              "      <td>7.2500</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>Cumings, Mrs. John Bradley (Florence Briggs Th...</td>\n",
              "      <td>0</td>\n",
              "      <td>38.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>PC 17599</td>\n",
              "      <td>71.2833</td>\n",
              "      <td>C85</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>Heikkinen, Miss. Laina</td>\n",
              "      <td>0</td>\n",
              "      <td>26.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>STON/O2. 3101282</td>\n",
              "      <td>7.9250</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>Futrelle, Mrs. Jacques Heath (Lily May Peel)</td>\n",
              "      <td>0</td>\n",
              "      <td>35.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>113803</td>\n",
              "      <td>53.1000</td>\n",
              "      <td>C123</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>Allen, Mr. William Henry</td>\n",
              "      <td>1</td>\n",
              "      <td>35.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>373450</td>\n",
              "      <td>8.0500</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>886</th>\n",
              "      <td>887</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>Montvila, Rev. Juozas</td>\n",
              "      <td>1</td>\n",
              "      <td>27.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>211536</td>\n",
              "      <td>13.0000</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>887</th>\n",
              "      <td>888</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>Graham, Miss. Margaret Edith</td>\n",
              "      <td>0</td>\n",
              "      <td>19.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>112053</td>\n",
              "      <td>30.0000</td>\n",
              "      <td>B42</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>888</th>\n",
              "      <td>889</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>Johnston, Miss. Catherine Helen \"Carrie\"</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>W./C. 6607</td>\n",
              "      <td>23.4500</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>889</th>\n",
              "      <td>890</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>Behr, Mr. Karl Howell</td>\n",
              "      <td>1</td>\n",
              "      <td>26.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>111369</td>\n",
              "      <td>30.0000</td>\n",
              "      <td>C148</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>890</th>\n",
              "      <td>891</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>Dooley, Mr. Patrick</td>\n",
              "      <td>1</td>\n",
              "      <td>32.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>370376</td>\n",
              "      <td>7.7500</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>891 rows × 12 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "     PassengerId  Survived  Pclass  ...     Fare  Cabin  Embarked\n",
              "0              1         0       3  ...   7.2500    NaN       0.0\n",
              "1              2         1       1  ...  71.2833    C85       1.0\n",
              "2              3         1       3  ...   7.9250    NaN       0.0\n",
              "3              4         1       1  ...  53.1000   C123       0.0\n",
              "4              5         0       3  ...   8.0500    NaN       0.0\n",
              "..           ...       ...     ...  ...      ...    ...       ...\n",
              "886          887         0       2  ...  13.0000    NaN       0.0\n",
              "887          888         1       1  ...  30.0000    B42       0.0\n",
              "888          889         0       3  ...  23.4500    NaN       0.0\n",
              "889          890         1       1  ...  30.0000   C148       1.0\n",
              "890          891         0       3  ...   7.7500    NaN       2.0\n",
              "\n",
              "[891 rows x 12 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 110
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yZI3wmAOFuJO"
      },
      "source": [
        "Y_train = train_data[['Survived']]\n",
        "X_train = train_data[['Pclass', 'Sex', 'Age', 'SibSp', 'Parch', 'Fare', 'Embarked']]\n",
        "X_train = np.nan_to_num(X_train)\n",
        "Y_train = np.nan_to_num(Y_train)\n",
        "\n",
        "Y_test = gender_submission[['Survived']]\n",
        "X_test = test_data[['Pclass', 'Sex', 'Age', 'SibSp', 'Parch', 'Fare', 'Embarked']]\n",
        "X_test = np.nan_to_num(X_test)\n",
        "Y_test = np.nan_to_num(Y_test)"
      ],
      "execution_count": 111,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k5BHLO0SKU_1"
      },
      "source": [
        "Y_train = np.array(Y_train)\n",
        "X_train = np.array(X_train)\n",
        "\n",
        "Y_test = np.array(Y_test)\n",
        "X_test = np.array(X_test)"
      ],
      "execution_count": 112,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "11MG4ER_KjDs"
      },
      "source": [
        "model = tf.keras.models.Sequential([\n",
        "    Dense(7, activation='relu'),\n",
        "    Dense(64, activation='relu'),\n",
        "    Dense(128, activation='relu'),\n",
        "    Dense(64, activation='relu'),\n",
        "    Dense(2, activation='softmax'),\n",
        "])"
      ],
      "execution_count": 113,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0iAtjTj3NBp9"
      },
      "source": [
        "model.compile(optimizer = tf.keras.optimizers.Adam(learning_rate= 0.001),\n",
        "              loss=tf.keras.losses.sparse_categorical_crossentropy,\n",
        "              metrics = ['accuracy'])"
      ],
      "execution_count": 114,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CTbEFM6QOG7H",
        "outputId": "47d1a589-913f-4db9-e579-2accd3a3d857"
      },
      "source": [
        "output = model.fit(X_train, Y_train, epochs = 200)"
      ],
      "execution_count": 115,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/200\n",
            "28/28 [==============================] - 1s 4ms/step - loss: 0.6625 - accuracy: 0.6308\n",
            "Epoch 2/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.6098 - accuracy: 0.6813\n",
            "Epoch 3/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.6133 - accuracy: 0.6891\n",
            "Epoch 4/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5895 - accuracy: 0.6958\n",
            "Epoch 5/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.6039 - accuracy: 0.6790\n",
            "Epoch 6/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5892 - accuracy: 0.6790\n",
            "Epoch 7/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5978 - accuracy: 0.6970\n",
            "Epoch 8/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5739 - accuracy: 0.6981\n",
            "Epoch 9/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5637 - accuracy: 0.7026\n",
            "Epoch 10/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.5653 - accuracy: 0.7082\n",
            "Epoch 11/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5669 - accuracy: 0.7093\n",
            "Epoch 12/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5621 - accuracy: 0.7194\n",
            "Epoch 13/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5713 - accuracy: 0.7071\n",
            "Epoch 14/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.5533 - accuracy: 0.7239\n",
            "Epoch 15/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5509 - accuracy: 0.7284\n",
            "Epoch 16/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5729 - accuracy: 0.7059\n",
            "Epoch 17/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.5533 - accuracy: 0.7160\n",
            "Epoch 18/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.5526 - accuracy: 0.7273\n",
            "Epoch 19/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5509 - accuracy: 0.7138\n",
            "Epoch 20/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5447 - accuracy: 0.7295\n",
            "Epoch 21/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5348 - accuracy: 0.7351\n",
            "Epoch 22/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5381 - accuracy: 0.7262\n",
            "Epoch 23/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5345 - accuracy: 0.7194\n",
            "Epoch 24/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5360 - accuracy: 0.7396\n",
            "Epoch 25/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5364 - accuracy: 0.7407\n",
            "Epoch 26/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5276 - accuracy: 0.7419\n",
            "Epoch 27/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.5220 - accuracy: 0.7464\n",
            "Epoch 28/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5208 - accuracy: 0.7430\n",
            "Epoch 29/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.5172 - accuracy: 0.7531\n",
            "Epoch 30/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.5092 - accuracy: 0.7497\n",
            "Epoch 31/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4961 - accuracy: 0.7643\n",
            "Epoch 32/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4889 - accuracy: 0.7699\n",
            "Epoch 33/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4957 - accuracy: 0.7722\n",
            "Epoch 34/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.4945 - accuracy: 0.7699\n",
            "Epoch 35/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4835 - accuracy: 0.7744\n",
            "Epoch 36/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4778 - accuracy: 0.7789\n",
            "Epoch 37/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.4709 - accuracy: 0.7778\n",
            "Epoch 38/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4696 - accuracy: 0.7834\n",
            "Epoch 39/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4645 - accuracy: 0.7811\n",
            "Epoch 40/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.4593 - accuracy: 0.7811\n",
            "Epoch 41/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4596 - accuracy: 0.7935\n",
            "Epoch 42/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4540 - accuracy: 0.7957\n",
            "Epoch 43/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4568 - accuracy: 0.7912\n",
            "Epoch 44/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4585 - accuracy: 0.7980\n",
            "Epoch 45/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4513 - accuracy: 0.7845\n",
            "Epoch 46/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4438 - accuracy: 0.7912\n",
            "Epoch 47/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4472 - accuracy: 0.7946\n",
            "Epoch 48/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4579 - accuracy: 0.7969\n",
            "Epoch 49/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4506 - accuracy: 0.7991\n",
            "Epoch 50/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4422 - accuracy: 0.8002\n",
            "Epoch 51/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4341 - accuracy: 0.8002\n",
            "Epoch 52/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4401 - accuracy: 0.7946\n",
            "Epoch 53/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4473 - accuracy: 0.7890\n",
            "Epoch 54/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.4280 - accuracy: 0.8092\n",
            "Epoch 55/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4376 - accuracy: 0.7991\n",
            "Epoch 56/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.4480 - accuracy: 0.8036\n",
            "Epoch 57/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4261 - accuracy: 0.8036\n",
            "Epoch 58/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4222 - accuracy: 0.8103\n",
            "Epoch 59/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4237 - accuracy: 0.8171\n",
            "Epoch 60/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4319 - accuracy: 0.8081\n",
            "Epoch 61/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4488 - accuracy: 0.7935\n",
            "Epoch 62/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4315 - accuracy: 0.8103\n",
            "Epoch 63/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.4204 - accuracy: 0.8013\n",
            "Epoch 64/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4136 - accuracy: 0.8103\n",
            "Epoch 65/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4098 - accuracy: 0.8193\n",
            "Epoch 66/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.4134 - accuracy: 0.8126\n",
            "Epoch 67/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4375 - accuracy: 0.8137\n",
            "Epoch 68/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4218 - accuracy: 0.8058\n",
            "Epoch 69/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4199 - accuracy: 0.8036\n",
            "Epoch 70/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4092 - accuracy: 0.8238\n",
            "Epoch 71/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4082 - accuracy: 0.8148\n",
            "Epoch 72/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4061 - accuracy: 0.8103\n",
            "Epoch 73/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4066 - accuracy: 0.8159\n",
            "Epoch 74/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.4018 - accuracy: 0.8316\n",
            "Epoch 75/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.4240 - accuracy: 0.8159\n",
            "Epoch 76/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4340 - accuracy: 0.8070\n",
            "Epoch 77/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4124 - accuracy: 0.8126\n",
            "Epoch 78/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3994 - accuracy: 0.8159\n",
            "Epoch 79/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4156 - accuracy: 0.8092\n",
            "Epoch 80/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4044 - accuracy: 0.8171\n",
            "Epoch 81/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3932 - accuracy: 0.8171\n",
            "Epoch 82/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3963 - accuracy: 0.8182\n",
            "Epoch 83/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4072 - accuracy: 0.8159\n",
            "Epoch 84/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4064 - accuracy: 0.8204\n",
            "Epoch 85/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3952 - accuracy: 0.8272\n",
            "Epoch 86/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3910 - accuracy: 0.8204\n",
            "Epoch 87/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3891 - accuracy: 0.8204\n",
            "Epoch 88/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3974 - accuracy: 0.8159\n",
            "Epoch 89/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3921 - accuracy: 0.8316\n",
            "Epoch 90/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3947 - accuracy: 0.8215\n",
            "Epoch 91/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.4009 - accuracy: 0.8204\n",
            "Epoch 92/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3955 - accuracy: 0.8193\n",
            "Epoch 93/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3965 - accuracy: 0.8103\n",
            "Epoch 94/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3930 - accuracy: 0.8171\n",
            "Epoch 95/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3899 - accuracy: 0.8137\n",
            "Epoch 96/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3881 - accuracy: 0.8204\n",
            "Epoch 97/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3876 - accuracy: 0.8272\n",
            "Epoch 98/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3789 - accuracy: 0.8215\n",
            "Epoch 99/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3905 - accuracy: 0.8182\n",
            "Epoch 100/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3918 - accuracy: 0.8126\n",
            "Epoch 101/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3902 - accuracy: 0.8103\n",
            "Epoch 102/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3833 - accuracy: 0.8249\n",
            "Epoch 103/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3915 - accuracy: 0.8159\n",
            "Epoch 104/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3935 - accuracy: 0.8294\n",
            "Epoch 105/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3865 - accuracy: 0.8227\n",
            "Epoch 106/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3856 - accuracy: 0.8215\n",
            "Epoch 107/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3848 - accuracy: 0.8283\n",
            "Epoch 108/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3804 - accuracy: 0.8305\n",
            "Epoch 109/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3794 - accuracy: 0.8260\n",
            "Epoch 110/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3802 - accuracy: 0.8227\n",
            "Epoch 111/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3794 - accuracy: 0.8283\n",
            "Epoch 112/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3828 - accuracy: 0.8227\n",
            "Epoch 113/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3898 - accuracy: 0.8227\n",
            "Epoch 114/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3804 - accuracy: 0.8227\n",
            "Epoch 115/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3744 - accuracy: 0.8238\n",
            "Epoch 116/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3767 - accuracy: 0.8114\n",
            "Epoch 117/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3998 - accuracy: 0.8137\n",
            "Epoch 118/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3910 - accuracy: 0.8137\n",
            "Epoch 119/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3779 - accuracy: 0.8328\n",
            "Epoch 120/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3721 - accuracy: 0.8328\n",
            "Epoch 121/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3689 - accuracy: 0.8316\n",
            "Epoch 122/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3739 - accuracy: 0.8361\n",
            "Epoch 123/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3740 - accuracy: 0.8339\n",
            "Epoch 124/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3738 - accuracy: 0.8283\n",
            "Epoch 125/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3771 - accuracy: 0.8361\n",
            "Epoch 126/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3791 - accuracy: 0.8260\n",
            "Epoch 127/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3660 - accuracy: 0.8395\n",
            "Epoch 128/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3660 - accuracy: 0.8339\n",
            "Epoch 129/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3620 - accuracy: 0.8339\n",
            "Epoch 130/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3713 - accuracy: 0.8373\n",
            "Epoch 131/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3713 - accuracy: 0.8238\n",
            "Epoch 132/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3793 - accuracy: 0.8294\n",
            "Epoch 133/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3756 - accuracy: 0.8283\n",
            "Epoch 134/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3677 - accuracy: 0.8339\n",
            "Epoch 135/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3687 - accuracy: 0.8238\n",
            "Epoch 136/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3631 - accuracy: 0.8283\n",
            "Epoch 137/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3677 - accuracy: 0.8361\n",
            "Epoch 138/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3702 - accuracy: 0.8350\n",
            "Epoch 139/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3756 - accuracy: 0.8272\n",
            "Epoch 140/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3561 - accuracy: 0.8339\n",
            "Epoch 141/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3619 - accuracy: 0.8294\n",
            "Epoch 142/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3602 - accuracy: 0.8418\n",
            "Epoch 143/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3622 - accuracy: 0.8350\n",
            "Epoch 144/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3564 - accuracy: 0.8339\n",
            "Epoch 145/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3548 - accuracy: 0.8373\n",
            "Epoch 146/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3605 - accuracy: 0.8406\n",
            "Epoch 147/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3653 - accuracy: 0.8418\n",
            "Epoch 148/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3552 - accuracy: 0.8440\n",
            "Epoch 149/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3586 - accuracy: 0.8361\n",
            "Epoch 150/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3811 - accuracy: 0.8294\n",
            "Epoch 151/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3696 - accuracy: 0.8350\n",
            "Epoch 152/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3643 - accuracy: 0.8260\n",
            "Epoch 153/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3702 - accuracy: 0.8361\n",
            "Epoch 154/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3659 - accuracy: 0.8339\n",
            "Epoch 155/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3592 - accuracy: 0.8339\n",
            "Epoch 156/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3552 - accuracy: 0.8474\n",
            "Epoch 157/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3594 - accuracy: 0.8361\n",
            "Epoch 158/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3542 - accuracy: 0.8418\n",
            "Epoch 159/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3710 - accuracy: 0.8294\n",
            "Epoch 160/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3514 - accuracy: 0.8462\n",
            "Epoch 161/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3541 - accuracy: 0.8395\n",
            "Epoch 162/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3465 - accuracy: 0.8429\n",
            "Epoch 163/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3531 - accuracy: 0.8429\n",
            "Epoch 164/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3487 - accuracy: 0.8451\n",
            "Epoch 165/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3526 - accuracy: 0.8328\n",
            "Epoch 166/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3517 - accuracy: 0.8474\n",
            "Epoch 167/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3558 - accuracy: 0.8429\n",
            "Epoch 168/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3426 - accuracy: 0.8507\n",
            "Epoch 169/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3471 - accuracy: 0.8519\n",
            "Epoch 170/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3540 - accuracy: 0.8406\n",
            "Epoch 171/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3520 - accuracy: 0.8462\n",
            "Epoch 172/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3594 - accuracy: 0.8339\n",
            "Epoch 173/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3438 - accuracy: 0.8418\n",
            "Epoch 174/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3536 - accuracy: 0.8406\n",
            "Epoch 175/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3422 - accuracy: 0.8496\n",
            "Epoch 176/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3529 - accuracy: 0.8339\n",
            "Epoch 177/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3414 - accuracy: 0.8552\n",
            "Epoch 178/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3440 - accuracy: 0.8519\n",
            "Epoch 179/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3508 - accuracy: 0.8552\n",
            "Epoch 180/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3434 - accuracy: 0.8552\n",
            "Epoch 181/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3460 - accuracy: 0.8451\n",
            "Epoch 182/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3554 - accuracy: 0.8440\n",
            "Epoch 183/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3479 - accuracy: 0.8530\n",
            "Epoch 184/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3468 - accuracy: 0.8530\n",
            "Epoch 185/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3433 - accuracy: 0.8462\n",
            "Epoch 186/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3407 - accuracy: 0.8541\n",
            "Epoch 187/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3438 - accuracy: 0.8496\n",
            "Epoch 188/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3396 - accuracy: 0.8507\n",
            "Epoch 189/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3480 - accuracy: 0.8451\n",
            "Epoch 190/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3508 - accuracy: 0.8440\n",
            "Epoch 191/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3525 - accuracy: 0.8507\n",
            "Epoch 192/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3492 - accuracy: 0.8462\n",
            "Epoch 193/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3365 - accuracy: 0.8429\n",
            "Epoch 194/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3379 - accuracy: 0.8586\n",
            "Epoch 195/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3452 - accuracy: 0.8597\n",
            "Epoch 196/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3369 - accuracy: 0.8597\n",
            "Epoch 197/200\n",
            "28/28 [==============================] - 0s 3ms/step - loss: 0.3382 - accuracy: 0.8552\n",
            "Epoch 198/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3348 - accuracy: 0.8519\n",
            "Epoch 199/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3491 - accuracy: 0.8429\n",
            "Epoch 200/200\n",
            "28/28 [==============================] - 0s 4ms/step - loss: 0.3335 - accuracy: 0.8530\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dECQhrSYNkDb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "b732ff30-21e8-47df-dcf3-69ff02c5e71b"
      },
      "source": [
        "plt.plot(output.history['loss'])"
      ],
      "execution_count": 116,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f4d39176110>]"
            ]
          },
          "metadata": {},
          "execution_count": 116
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OjYpVrVeOdCe",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "af56523d-2b7d-4a09-b17c-9d43e0b91d08"
      },
      "source": [
        "plt.plot(output.history['accuracy'])"
      ],
      "execution_count": 117,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f4d47706b90>]"
            ]
          },
          "metadata": {},
          "execution_count": 117
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "89ppiBExdh9Z"
      },
      "source": [
        "class KNearestNeighbors:\n",
        "    def __init__(self, k):\n",
        "        self.k = k\n",
        "        \n",
        "    #train\n",
        "    def fit(self, X_train, Y_train):\n",
        "        self.X_train = X_train\n",
        "        self.Y_train = Y_train\n",
        "        self.number_class = len(np.unique(Y_train))\n",
        "        \n",
        "    def nearestNeighbors(self, X_test):\n",
        "        distance = np.sqrt(np.sum((X_test - self.X_train)**2, axis = 1))\n",
        "        near_neighbor = np.argsort(distance)[0:self.k]\n",
        "        return near_neighbor\n",
        "    \n",
        "    #test\n",
        "    def predict(self, X_test):\n",
        "        near_neighbor = self.nearestNeighbors(X_test)\n",
        "        y = np.argmax(np.bincount(self.Y_train[near_neighbor]))\n",
        "        return y\n",
        "    \n",
        "    def evaluate(self, x_test, y_test):\n",
        "        y_pred = []\n",
        "        for i in range(len(x_test)):\n",
        "          y = knn.predict(x_test[i])\n",
        "          y_pred.append(y)\n",
        "        \n",
        "        not_correct = abs(sum(y_pred - y_test))\n",
        "        return ((((len(y_pred) - not_correct)*100)/len(y_pred))/100)     "
      ],
      "execution_count": 119,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9AYoHdY4uCOf",
        "outputId": "ccda57f7-6358-438b-efd8-239c9293dbd4"
      },
      "source": [
        "Y_train = Y_train.reshape(-1)\n",
        "Y_test = Y_test.reshape(-1)\n",
        "knn = KNearestNeighbors(5)\n",
        "knn.fit(X_train, Y_train)\n",
        "\n",
        "print('k= 5  Accuracy=', knn.evaluate(X_test, Y_test))"
      ],
      "execution_count": 120,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "k= 5  Accuracy= 0.9904306220095694\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N10Y7gnIua8o"
      },
      "source": [
        "class AdalineClassifier:\n",
        "    def __init__(self):\n",
        "        pass\n",
        "    \n",
        "    def fit(self, X_train, Y_train):\n",
        "        self.w = np.matmul(inv(np.matmul(X_train.T, X_train)), np.matmul(X_train.T, Y_train))\n",
        "\n",
        "    def predict(self, X_test):\n",
        "        Y_predic = np.matmul(X_test, self.w)\n",
        "        return Y_predic\n",
        "    \n",
        "    def evaluation(self, X_test, Y_test):\n",
        "        Y_predic = np.matmul(X_test, self.w)\n",
        "        subtract = np.abs(Y_test - Y_predic)\n",
        "        \n",
        "        correct = 0\n",
        "        for i in range(len(Y_pred)):\n",
        "          if Y_test[i] == 0 and Y_pred[i] < 0.5:\n",
        "            correct += 1\n",
        "          if Y_test[i] == 1 and Y_pred[i] > 0.5:\n",
        "            correct += 1\n",
        "\n",
        "        average = np.mean(subtract)\n",
        "        acc = (correct * 100) / len(Y_test) / 100\n",
        "        return average, acc\n"
      ],
      "execution_count": 121,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "R6p1j27Tva_S",
        "outputId": "b621cf66-cee1-4f22-9ea5-6865f109b1f0"
      },
      "source": [
        "model = AdalineClassifier()\n",
        "model.fit(X_train, Y_train)\n",
        "Y_pred = model.predict(X_test)\n",
        "result = model.evaluation(X_test, Y_test)\n",
        "print('evaluation= ',result[0],' Accuracy= ',result[1])"
      ],
      "execution_count": 122,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "evaluation=  0.2731463040521892  Accuracy=  0.8971291866028709\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bFg98IqG0Hhl",
        "outputId": "47d6da35-f4f2-4366-ca95-865398fc151c"
      },
      "source": [
        "print(X_train.shape, Y_train.shape)\n",
        "print(X_test.shape, Y_test.shape)"
      ],
      "execution_count": 123,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(891, 7) (891,)\n",
            "(418, 7) (418,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MTQwyDQivmzw"
      },
      "source": [
        "class Perceptron:\n",
        "    def __init__(self):\n",
        "        self.W = np.random.rand(7)\n",
        "        self.b = np.random.rand(1)\n",
        "        self.lr = 0.00001\n",
        "        self.loss = []\n",
        "\n",
        "    def fit(self, x_train, y_train):\n",
        "\n",
        "        for i in range(len(x_train)):    \n",
        "            y_pred = np.matmul(x_train[i], self.W) + self.b\n",
        "            e = y_train[i] - y_pred\n",
        "            self.W = self.W + e * self.lr * x_train[i]\n",
        "            self.b = self.b + e * self.lr\n",
        "\n",
        "            Y_pred = np.matmul(X_train, self.W)\n",
        "            error = np.mean(np.abs(Y_train - Y_pred))\n",
        "            self.loss.append(error)\n",
        "\n",
        "        return self.loss\n",
        "\n",
        "    def predict(self, x):\n",
        "        x = np.array(x)\n",
        "        Y_pred = np.matmul(x, self.W) + self.b\n",
        "        return Y_pred\n",
        "\n",
        "\n",
        "    def evaluation(self, X, Y):\n",
        "        Y_predic = np.matmul(X, self.W) + self.b\n",
        "        correct = 0\n",
        "        for i in range(len(Y_predic)):\n",
        "          if Y[i] == 0 and Y_predic[i] < 0.5:\n",
        "            correct += 1\n",
        "          if Y[i] == 1 and Y_predic[i] > 0.5:\n",
        "            correct += 1\n",
        "\n",
        "        acc = (correct * 100) / len(Y) / 100\n",
        "        return acc"
      ],
      "execution_count": 124,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UB4z9RFgxj4l"
      },
      "source": [
        "Y_train = Y_train.reshape(-1,1)\n",
        "perceptron_model = Perceptron()\n",
        "loss = perceptron_model.fit(X_train, Y_train)"
      ],
      "execution_count": 125,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZaT8hi5WzLBd",
        "outputId": "9816bab3-8c86-404b-f550-35d013eed7ee"
      },
      "source": [
        "Y_pred = perceptron_model.predict(X_test)\n",
        "\n",
        "Y_test = Y_test.reshape(-1,1)\n",
        "evaluate = perceptron_model.evaluation(X_test, Y_test)\n",
        "print('test data accuracy: ', evaluate)"
      ],
      "execution_count": 126,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "test data accuracy:  0.4138755980861244\n"
          ]
        }
      ]
    }
  ]
}