Assignment.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Y5y_HbJiPKhA"
      },
      "source": [
        "# Linear Regression\n",
        "## Question 1\n",
        "Make a class called LinearRegression which provides two functions : fit and predict. Try to implement it from scratch. If stuck, refer to the examples folder."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 275,
      "metadata": {
        "id": "gzoG2XilPLFr"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "class LinearRegression():\n",
        "    \n",
        "    def __init__(self, iterations, learning_rate):\n",
        "        self.learning_rate = learning_rate\n",
        "        self.iterations = iterations\n",
        "\n",
        "    def fit(self, X_train, y_train):\n",
        "        self.weights = np.zeros(X_train[0].shape)\n",
        "        self.bias = 0\n",
        "        m = len(y_train)\n",
        "        self.costs = []\n",
        "        for _ in range(self.iterations):\n",
        "            y_pred = X_train.dot(self.weights) + self.bias\n",
        "            # Loss in weights = XT(Xw +b - y)\n",
        "            dw = (X_train.T.dot(y_pred - y_train))/m\n",
        "            self.weights -= self.learning_rate*dw\n",
        "            # append the cost in each iteration\n",
        "            self.costs.append(np.sum(np.square(y_pred - y_train)))\n",
        "            # Loss in bias in batch GD = sum of (yhat - y) for each training example\n",
        "            db = (np.sum(y_pred - y_train))/m\n",
        "            self.bias -= self.learning_rate*db\n",
        "    def predict(self, X_test):\n",
        "        n = len(X_test)\n",
        "        predictions = []\n",
        "        for idx in range(n):\n",
        "            yhat = self.weights.dot(X_test[idx]) + self.bias\n",
        "            predictions.append(yhat)\n",
        "        return predictions\n",
        "    def cost(self):\n",
        "        return self.costs\n",
        "            \n",
        "        "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PsqoxNag7D3-"
      },
      "source": [
        "## Question 2\n",
        "\n",
        "Use the dataset https://www.kaggle.com/datasets/quantbruce/real-estate-price-prediction (*).\n",
        "1. Read it using pandas.\n",
        "2. Check for **null values**.\n",
        "3. For each of the columns (except the first and last), plot the column values in the X-axis against the last column of prices in the Y-axis.\n",
        "4. Remove the unwanted columns.\n",
        "5. Split the dataset into train and test data. Test data size = 25% of total dataset.\n",
        "6. **Normalize** the X_train and X_test using MinMaxScaler from sklearn.preprocessing.\n",
        "7. Fit the training data into the model created in question 1 and predict the testing data.\n",
        "8. Use **mean square error and R<sup>2</sup>** from sklearn.metrics as evaluation criterias.\n",
        "9. Fit the training data into the models of the same name provided by sklearn.linear_model and evaluate the predictions using MSE and R<sup>2</sup>.\n",
        "10. Tune the hyperparameters of your models (learning rate, epochs) to achieve losses close to that of the sklearn models.\n",
        "\n",
        "Note : (*) To solve this question, you may proceed in any of the following ways :\n",
        "1. Prepare the notebook in Kaggle, download it and submit it separately with the other questions.\n",
        "2. Download the dataset from kaggle. Upload it to the session storage in Colab.\n",
        "3. Use Colab data directly in Colab. [Refer here](https://www.kaggle.com/general/74235). For this, you need to create kaggle API token. Before submitting, hide or remove the API token."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 276,
      "metadata": {
        "id": "8lupaMcr63QF"
      },
      "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>No</th>\n",
              "      <th>X1 transaction date</th>\n",
              "      <th>X2 house age</th>\n",
              "      <th>X3 distance to the nearest MRT station</th>\n",
              "      <th>X4 number of convenience stores</th>\n",
              "      <th>X5 latitude</th>\n",
              "      <th>X6 longitude</th>\n",
              "      <th>Y house price of unit area</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>2012.917</td>\n",
              "      <td>32.0</td>\n",
              "      <td>84.87882</td>\n",
              "      <td>10</td>\n",
              "      <td>24.98298</td>\n",
              "      <td>121.54024</td>\n",
              "      <td>37.9</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>2012.917</td>\n",
              "      <td>19.5</td>\n",
              "      <td>306.59470</td>\n",
              "      <td>9</td>\n",
              "      <td>24.98034</td>\n",
              "      <td>121.53951</td>\n",
              "      <td>42.2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>2013.583</td>\n",
              "      <td>13.3</td>\n",
              "      <td>561.98450</td>\n",
              "      <td>5</td>\n",
              "      <td>24.98746</td>\n",
              "      <td>121.54391</td>\n",
              "      <td>47.3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>2013.500</td>\n",
              "      <td>13.3</td>\n",
              "      <td>561.98450</td>\n",
              "      <td>5</td>\n",
              "      <td>24.98746</td>\n",
              "      <td>121.54391</td>\n",
              "      <td>54.8</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5</td>\n",
              "      <td>2012.833</td>\n",
              "      <td>5.0</td>\n",
              "      <td>390.56840</td>\n",
              "      <td>5</td>\n",
              "      <td>24.97937</td>\n",
              "      <td>121.54245</td>\n",
              "      <td>43.1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   No  X1 transaction date  X2 house age  \\\n",
              "0   1             2012.917          32.0   \n",
              "1   2             2012.917          19.5   \n",
              "2   3             2013.583          13.3   \n",
              "3   4             2013.500          13.3   \n",
              "4   5             2012.833           5.0   \n",
              "\n",
              "   X3 distance to the nearest MRT station  X4 number of convenience stores  \\\n",
              "0                                84.87882                               10   \n",
              "1                               306.59470                                9   \n",
              "2                               561.98450                                5   \n",
              "3                               561.98450                                5   \n",
              "4                               390.56840                                5   \n",
              "\n",
              "   X5 latitude  X6 longitude  Y house price of unit area  \n",
              "0     24.98298     121.54024                        37.9  \n",
              "1     24.98034     121.53951                        42.2  \n",
              "2     24.98746     121.54391                        47.3  \n",
              "3     24.98746     121.54391                        54.8  \n",
              "4     24.97937     121.54245                        43.1  "
            ]
          },
          "execution_count": 276,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "import os\n",
        "import pandas as pd\n",
        "os.chdir(os.path.expanduser('~/Downloads'))\n",
        "realestate = pd.read_csv('Real estate.csv')\n",
        "realestate.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 277,
      "metadata": {},
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 278,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<Axes: xlabel='X4 number of convenience stores', ylabel='Y house price of unit area'>"
            ]
          },
          "execution_count": 278,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 200x200 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.figure(figsize=(2, 2))\n",
        "sns.barplot(x = 'X4 number of convenience stores', y = 'Y house price of unit area', data=realestate)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 279,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "No                                        False\n",
              "X1 transaction date                       False\n",
              "X2 house age                              False\n",
              "X3 distance to the nearest MRT station    False\n",
              "X4 number of convenience stores           False\n",
              "X5 latitude                               False\n",
              "X6 longitude                              False\n",
              "Y house price of unit area                False\n",
              "dtype: bool"
            ]
          },
          "execution_count": 279,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "realestate.isnull().any()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 280,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1000x200 with 5 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "columns = ['X2 house age',\t'X3 distance to the nearest MRT station', 'X4 number of convenience stores', 'X5 latitude', \n",
        "           'X6 longitude']\n",
        "fig, axs = plt.subplots(1, 5, figsize=(10, 2))\n",
        "yplot = 'Y house price of unit area'\n",
        "axs[0].set_ylabel('house price')\n",
        "for i, feature in enumerate(columns):\n",
        "    if feature == 'X4 number of convenience stores':\n",
        "        axs[i].bar(realestate[feature], realestate[yplot])\n",
        "    axs[i].set_xlabel(f'X{i+2}')\n",
        "    axs[i].scatter(x=realestate[feature], y=realestate[yplot])\n",
        "\n",
        "    "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 281,
      "metadata": {},
      "outputs": [],
      "source": [
        "from sklearn.preprocessing import MinMaxScaler, Normalizer\n",
        "\n",
        "X = np.array(realestate.drop(columns=['No', 'X1 transaction date']))\n",
        "y = np.array(realestate['Y house price of unit area'])\n",
        "\n",
        "minmaxscalerX = MinMaxScaler()\n",
        "standardscalery = Normalizer()\n",
        "\n",
        "X = minmaxscalerX.fit_transform(X)\n",
        "y = standardscalery.fit_transform(y.reshape(1, -1))[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 282,
      "metadata": {},
      "outputs": [],
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 283,
      "metadata": {},
      "outputs": [],
      "source": [
        "lgr = LinearRegression(iterations=10000, learning_rate=0.01)\n",
        "lgr.fit(X_train, y_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 284,
      "metadata": {},
      "outputs": [],
      "source": [
        "y_predicted = lgr.predict(X_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 285,
      "metadata": {},
      "outputs": [],
      "source": [
        "from sklearn.metrics import mean_squared_error\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 286,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "Text(0, 0.002, 'cost = 0.008388613588750666')"
            ]
          },
          "execution_count": 286,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 300x300 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "cost = lgr.cost()\n",
        "epochs = np.arange(lgr.iterations)\n",
        "plt.figure(figsize=(3, 3))\n",
        "plt.plot(epochs, cost)\n",
        "plt.xlabel('epoch')\n",
        "plt.ylabel('cost')\n",
        "plt.title('plot of cost in each iteration/epoch')\n",
        "# Cost at last epoch .\n",
        "plt.text(0, 0.0020, f'cost = {cost[-1]}')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 287,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Accuracy based on scratch model : 2.2259114195413127e-05\n",
            "based on Sckit learn Linear regression 1.755703894765389e-33\n"
          ]
        }
      ],
      "source": [
        "# Sklearn Model Linear regression\n",
        "from sklearn.linear_model import LinearRegression\n",
        "\n",
        "lnrrgr = LinearRegression(n_jobs=-1)\n",
        "lnrrgr.fit(X_train, y_train)\n",
        "y_predicted_sk = lnrrgr.predict(X_test)\n",
        "print(f'Accuracy based on scratch model : {mean_squared_error(y_predicted, y_test)}'); \n",
        "print(f'based on Sckit learn Linear regression {mean_squared_error(y_test, y_predicted_sk)}'); "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dQ7lQpy-SYCq"
      },
      "source": [
        "# Logistic Regression\n",
        "## Question 3\n",
        "\n",
        "The breast cancer dataset is a binary classification dataset commonly used in machine learning tasks. It is available in scikit-learn (sklearn) as part of its datasets module.\n",
        "Here is an explanation of the breast cancer dataset's components:\n",
        "\n",
        "* Features (X):\n",
        "\n",
        " * The breast cancer dataset consists of 30 numeric features representing different characteristics of the FNA images. These features include mean, standard error, and worst (largest) values of various attributes such as radius, texture, smoothness, compactness, concavity, symmetry, fractal dimension, etc.\n",
        "\n",
        "* Target (y):\n",
        "\n",
        " * The breast cancer dataset is a binary classification problem, and the target variable (y) represents the diagnosis of the breast mass. It contains two classes:\n",
        "    * 0: Represents a malignant (cancerous) tumor.\n",
        "    * 1: Represents a benign (non-cancerous) tumor.\n",
        "\n",
        "Complete the code given below in place of the \"...\""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "auipk-zBpmO-"
      },
      "source": [
        "1. Load the dataset from sklearn.datasets\n",
        "2. Separate out the X and Y columns.\n",
        "3. Normalize the X data using MinMaxScaler or StandardScaler.\n",
        "4. Create a train-test-split. Take any suitable test size."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 248,
      "metadata": {
        "id": "0OyGNHNjFh13"
      },
      "outputs": [],
      "source": [
        "from sklearn import datasets\n",
        "\n",
        "data  = datasets.load_breast_cancer()\n",
        "Xl = data['data'] ; yl = data['target']\n",
        "minmaxscalerXl = MinMaxScaler()\n",
        "Xl = minmaxscalerXl.fit_transform(Xl)\n",
        "\n",
        "Xl_train, Xl_test, yl_train, yl_test = train_test_split(Xl, yl)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uM-SsSxpqF2o"
      },
      "source": [
        "5. Write code for the sigmoid function and Logistic regression.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 249,
      "metadata": {
        "id": "o81LA5MZFoTW"
      },
      "outputs": [],
      "source": [
        "def sigmoid(z):\n",
        "  g_z = 1/(1 + np.exp(-z))\n",
        "  return g_z\n",
        "\n",
        "def sigmoid_derivative(z):\n",
        "    return sigmoid(z) * (1 - sigmoid(z))\n",
        "\n",
        "class LogisticRegression:\n",
        "    def __init__(self, learning_rate, epochs):\n",
        "      #Initialise the hyperparameters of the model\n",
        "        self.lr = learning_rate\n",
        "        self.epochs = epochs\n",
        "\n",
        "    def fit(self, X, y):\n",
        "        n_samples, n_features = X.shape\n",
        "        y = y.reshape(-1, 1)\n",
        "        self.weights = np.zeros((n_features, 1))\n",
        "        self.bias = 0\n",
        "\n",
        "        #Implement the GD algortihm\n",
        "        for _ in range(self.epochs):\n",
        "            z = X.dot(self.weights) + (self.bias)\n",
        "            y_pred = sigmoid(z)\n",
        "            dw = (X.T.dot(y_pred - y)/n_samples)\n",
        "            db = (np.sum(y_pred - y))/n_samples\n",
        "\n",
        "            self.weights -= self.lr * dw\n",
        "            self.bias -= self.lr * db\n",
        "\n",
        "    def predict(self, X):\n",
        "      #Write the predict function\n",
        "        z = X.dot(self.weights) + self.bias\n",
        "        y_pred = sigmoid(z)\n",
        "          \n",
        "          \n",
        "        return y_pred"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Uo9LNRMzq4K-"
      },
      "source": [
        "6. Fit your model on the dataset and make predictions.\n",
        "7. Compare your model with the Sklearn Logistic Regression model. Try out all the different penalties.\n",
        "8. Print accuracy_score in each case using sklearn.metrics ."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 250,
      "metadata": {
        "id": "DyGsTTOqFphf"
      },
      "outputs": [],
      "source": [
        "logreg = LogisticRegression(epochs=500, learning_rate=0.01)\n",
        "logreg.fit(Xl_train, yl_train)\n",
        "yl_predict = logreg.predict(Xl_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 251,
      "metadata": {},
      "outputs": [],
      "source": [
        "from sklearn.linear_model import LogisticRegression\n",
        "\n",
        "logit = LogisticRegression()\n",
        "logit.fit(Xl_train, yl_train)\n",
        "yl_predict_sk = logit.predict(Xl_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 253,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x15eb66d50>"
            ]
          },
          "execution_count": 253,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 700x300 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "from sklearn.metrics import roc_curve, roc_auc_score\n",
        "\n",
        "fpr, tpr, thresholds = roc_curve(yl_test, yl_predict)\n",
        "fig, axs = plt.subplots(1, 2, figsize = (7, 3))\n",
        "plt.title('based on scratch model')\n",
        "axs[0].plot(fpr, tpr)\n",
        "axs[0].set_xlim(-0.02, 1.0)\n",
        "axs[0].set_ylim(0, 1.1)\n",
        "axs[0].plot([0, 1], [0, 1], 'r--')\n",
        "axs[1].plot(thresholds, tpr, label='tpr')\n",
        "axs[1].plot(thresholds, fpr, 'g', label='fpr')\n",
        "axs[1].legend()\n",
        "# Best threshold is 0.52\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 254,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x15ebf5d90>"
            ]
          },
          "execution_count": 254,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 700x300 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "from sklearn.metrics import roc_curve, roc_auc_score\n",
        "\n",
        "fpr, tpr, thresholds = roc_curve(yl_test, yl_predict_sk)\n",
        "fig, axs = plt.subplots(1, 2, figsize = (7, 3))\n",
        "plt.title('based on scikit learn model')\n",
        "axs[0].plot(fpr, tpr)\n",
        "axs[0].set_xlim(-0.02, 1.0)\n",
        "axs[0].set_ylim(0, 1.1)\n",
        "axs[0].plot([0, 1], [0, 1], 'r--')\n",
        "axs[1].plot(thresholds, tpr, label='tpr')\n",
        "axs[1].plot(thresholds, fpr, 'g', label='fpr')\n",
        "axs[1].set_xlim(-0.02, 1.1)\n",
        "axs[1].legend()\n",
        "# Best threshold is 0.9"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AGBkzAO5red4"
      },
      "source": [
        "9. For the best model in each case (yours and scikit-learn), print the classification_report using sklearn.metrics .\n",
        "10. For the best model in each case (yours and scikit-learn), print the confusion_matrix using sklearn.metrics ."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 260,
      "metadata": {},
      "outputs": [],
      "source": [
        "yl_predicted = []\n",
        "yl_predicted_sk = []\n",
        "for p_1, p_2 in zip(yl_predict, yl_predict_sk):\n",
        "    if p_1 >= 0.55 :\n",
        "        yl_predicted.append(1)\n",
        "    else: \n",
        "        yl_predicted.append(0)\n",
        "    if p_2 >= 0.95:\n",
        "        yl_predicted_sk.append(1)\n",
        "    else : \n",
        "        yl_predicted_sk.append(0)\n",
        "    "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 261,
      "metadata": {
        "id": "le-HfABsvnyF"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.89      0.94      0.92        52\n",
            "           1       0.97      0.93      0.95        91\n",
            "\n",
            "    accuracy                           0.94       143\n",
            "   macro avg       0.93      0.94      0.93       143\n",
            "weighted avg       0.94      0.94      0.94       143\n",
            "\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import classification_report\n",
        "print(classification_report(yl_test, yl_predicted))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 262,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      0.88      0.94        52\n",
            "           1       0.94      1.00      0.97        91\n",
            "\n",
            "    accuracy                           0.96       143\n",
            "   macro avg       0.97      0.94      0.95       143\n",
            "weighted avg       0.96      0.96      0.96       143\n",
            "\n"
          ]
        }
      ],
      "source": [
        "print(classification_report(yl_test, yl_predicted_sk))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 263,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'based on scratch model')"
            ]
          },
          "execution_count": 263,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "from sklearn.metrics import confusion_matrix\n",
        "sns.heatmap(confusion_matrix(y_true=yl_test, y_pred=yl_predicted), annot=True, cmap='Reds')\n",
        "plt.title('based on scratch model')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 264,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'based on scikit learn model')"
            ]
          },
          "execution_count": 264,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "sns.heatmap(confusion_matrix(y_true=yl_test, y_pred=yl_predicted_sk), annot=True, cmap='Reds')\n",
        "plt.title('based on scikit learn model')\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6OQ2tSp0MO6n"
      },
      "source": [
        "# KNN\n",
        "## Question 4\n",
        "\n",
        "How accurately can a K-Nearest Neighbors (KNN) model classify different types of glass based on a glass classification dataset consisting of 214 samples and 7 classes? Use the kaggle dataset \"https://www.kaggle.com/datasets/uciml/glass\".\n",
        "\n",
        "Context: This is a Glass Identification Data Set from UCI. It contains 10 attributes including id. The response is glass type(discrete 7 values)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iMGxbtX-zfsI"
      },
      "source": [
        "1. Load the data as you did in the 2nd question.\n",
        "2. Extract the X and Y columns.\n",
        "3. Split it into training and testing datasets."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 288,
      "metadata": {
        "id": "p0SfLB7pO7_z"
      },
      "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>RI</th>\n",
              "      <th>Na</th>\n",
              "      <th>Mg</th>\n",
              "      <th>Al</th>\n",
              "      <th>Si</th>\n",
              "      <th>K</th>\n",
              "      <th>Ca</th>\n",
              "      <th>Ba</th>\n",
              "      <th>Fe</th>\n",
              "      <th>Type</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1.52101</td>\n",
              "      <td>13.64</td>\n",
              "      <td>4.49</td>\n",
              "      <td>1.10</td>\n",
              "      <td>71.78</td>\n",
              "      <td>0.06</td>\n",
              "      <td>8.75</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1.51761</td>\n",
              "      <td>13.89</td>\n",
              "      <td>3.60</td>\n",
              "      <td>1.36</td>\n",
              "      <td>72.73</td>\n",
              "      <td>0.48</td>\n",
              "      <td>7.83</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1.51618</td>\n",
              "      <td>13.53</td>\n",
              "      <td>3.55</td>\n",
              "      <td>1.54</td>\n",
              "      <td>72.99</td>\n",
              "      <td>0.39</td>\n",
              "      <td>7.78</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1.51766</td>\n",
              "      <td>13.21</td>\n",
              "      <td>3.69</td>\n",
              "      <td>1.29</td>\n",
              "      <td>72.61</td>\n",
              "      <td>0.57</td>\n",
              "      <td>8.22</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1.51742</td>\n",
              "      <td>13.27</td>\n",
              "      <td>3.62</td>\n",
              "      <td>1.24</td>\n",
              "      <td>73.08</td>\n",
              "      <td>0.55</td>\n",
              "      <td>8.07</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        RI     Na    Mg    Al     Si     K    Ca   Ba   Fe  Type\n",
              "0  1.52101  13.64  4.49  1.10  71.78  0.06  8.75  0.0  0.0     1\n",
              "1  1.51761  13.89  3.60  1.36  72.73  0.48  7.83  0.0  0.0     1\n",
              "2  1.51618  13.53  3.55  1.54  72.99  0.39  7.78  0.0  0.0     1\n",
              "3  1.51766  13.21  3.69  1.29  72.61  0.57  8.22  0.0  0.0     1\n",
              "4  1.51742  13.27  3.62  1.24  73.08  0.55  8.07  0.0  0.0     1"
            ]
          },
          "execution_count": 288,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "glassdf = pd.read_csv('glass.csv.xls')\n",
        "glassdf.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 218,
      "metadata": {},
      "outputs": [],
      "source": [
        "Xk = np.array(glassdf.drop(columns=['RI', 'Type']))\n",
        "yk = np.array(glassdf.Type)\n",
        "Xk_train, Xk_test, yk_train, yk_test = train_test_split(Xk, yk, test_size=0.01)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qtyZJXh9zoh5"
      },
      "source": [
        "4. Define Euclidean distance.\n",
        "5. Build the KNN model.\n",
        "6. Fit the model on the training data. (Note : you may require to change the type of the data from pandas dataframe to numpy arrays. To do that, just do this X=np.array(X) and so on...)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 219,
      "metadata": {
        "id": "YJkhLORLzn6r"
      },
      "outputs": [],
      "source": [
        "def Euclidean_distance(X, y):\n",
        "    dis = np.square(X - y)\n",
        "    dis = np.sum(dis, axis=1)\n",
        "    return dis\n",
        "\n",
        "class KNN():\n",
        "    def __init__(self, k) :\n",
        "        # k nearest neighbours are considered.\n",
        "        self.k = k\n",
        "    def train(self, X, y):\n",
        "        self.X = X\n",
        "        self.y = y\n",
        "    def predict(self, X_test):\n",
        "        predictions = []\n",
        "        for x in X_test:\n",
        "            dis = Euclidean_distance(self.X , x)\n",
        "            dis = np.array(dis)\n",
        "            arg_dis = np.argsort(dis)[:self.k]\n",
        "            # argdis = []\n",
        "            # for idx in arg_dis:\n",
        "            #     argdis.append(self.y[idx])\n",
        "            # hash = {i : 0 for i in argdis}\n",
        "            # for class_ in argdis:\n",
        "            #     hash[class_]+=1\n",
        "            # for idx, (key, value) in enumerate(hash.items()):\n",
        "            #     if idx == 0:\n",
        "            #         maxkey = key\n",
        "            #         maxvalue = value\n",
        "            #     else:\n",
        "            #         if value > maxvalue :\n",
        "            #             maxvalue = value\n",
        "            #             maxkey = key\n",
        "            # predictions.append(maxkey)\n",
        "            nearest_labels = self.y[arg_dis]\n",
        "            \n",
        "            counts = np.bincount(nearest_labels)\n",
        "            prediction = np.argmax(counts)\n",
        "            \n",
        "            predictions.append(prediction)\n",
        "        \n",
        "        return predictions\n",
        "def accuracy(predictions,y_test):\n",
        "    return np.sum(predictions==y_test)/len(y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 220,
      "metadata": {},
      "outputs": [],
      "source": [
        "knn = KNN(k=3)\n",
        "knn.train(Xk_train, yk_train)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E9rxZpPB0pVS"
      },
      "source": [
        "7. Make predictions. Find their accuracy using accuracy_score. Try different k values. k=3 worked well in our case.\n",
        "8. Compare with the sklearn model (from sklearn.neighbors import KNeighborsClassifier)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 221,
      "metadata": {
        "id": "ernfjaZJ0pAh"
      },
      "outputs": [],
      "source": [
        "yk_predict = knn.predict(Xk_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 222,
      "metadata": {
        "id": "h2tZQg4L09wn"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "1.0"
            ]
          },
          "execution_count": 222,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# from sklearn.metrics import accuracy_score\n",
        "accuracy(yk_test, yk_predict)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 223,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "1.0"
            ]
          },
          "execution_count": 223,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "knn2 = KNN(k=5)\n",
        "knn2.train(Xk_train, yk_train)\n",
        "yk2_predict = knn2.predict(Xk_test)\n",
        "accuracy(yk_test, yk2_predict)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 224,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "1.0"
            ]
          },
          "execution_count": 224,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "knn3 = KNN(k=7)\n",
        "knn3.train(Xk_train, yk_train)\n",
        "yk3_predict = knn3.predict(Xk_test)\n",
        "accuracy(yk_test, yk3_predict)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 225,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "0.6666666666666666"
            ]
          },
          "execution_count": 225,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "knn4 = KNN(k=9)\n",
        "knn4.train(Xk_train, yk_train)\n",
        "yk4_predict = knn4.predict(Xk_test)\n",
        "accuracy(yk_test, yk4_predict)"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "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": "3.11.5"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}