newKNN.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "newKNN.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyOuiHjztLsz1yLjYDZUytcW",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Yuyan-C/MAIS202Project/blob/main/newKNN.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1odVAfcqe9oe",
        "outputId": "7440695a-ec29-4838-86ea-d162c42f680c"
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from datetime import datetime\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.linear_model import SGDClassifier\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "from sklearn.metrics import balanced_accuracy_score\n",
        "from sklearn.metrics import accuracy_score\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.metrics import classification_report\n",
        "from sklearn.svm import SVC\n",
        "import seaborn as sns\n",
        "from imblearn.over_sampling import RandomOverSampler, SMOTE, ADASYN\n",
        "df = pd.read_csv('covtype.csv')\n",
        "x = df['Horizontal_Distance_To_Hydrology'].values\n",
        "y = df['Vertical_Distance_To_Hydrology'].values\n",
        "distance = np.sqrt(np.square(x)+np.square(y))\n",
        "df = df.drop(['Horizontal_Distance_To_Hydrology', 'Vertical_Distance_To_Hydrology'], axis=1)\n",
        "df.insert(3, \"Distance_To_Hydrology\", distance)\n",
        "# set the number of training set, validation set, and testing set\n",
        "# get the number of features \n",
        "num_features = len(df.columns)-1\n",
        "# set X and y\n",
        "scaler = MinMaxScaler() \n",
        "X = scaler.fit_transform(df.iloc[:, :num_features])\n",
        "y = df.iloc[:, num_features:].to_numpy().reshape(len(df.index))\n",
        "# set training, validation, and testing sets\n",
        "# 64% training, 16% validation, 20% testing\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, shuffle = True)\n",
        "print(\"testing set:\", y_test.shape)\n",
        "X_train, X_valid, y_train, y_valid = train_test_split(X_train, y_train, test_size=0.2, random_state=42, shuffle = True)\n",
        "print(\"validation set:\", y_valid.shape)\n",
        "ros = RandomOverSampler(random_state=0)\n",
        "X_train, y_train = ros.fit_resample(X_train, y_train)\n",
        "print(\"training set:\", y_train.shape)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "testing set: (116203,)\n",
            "validation set: (92962,)\n",
            "training set: (1270115,)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/utils/deprecation.py:87: FutureWarning: Function safe_indexing is deprecated; safe_indexing is deprecated in version 0.22 and will be removed in version 0.24.\n",
            "  warnings.warn(msg, category=FutureWarning)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "oV4jR7qWfpJX",
        "outputId": "998194c6-d01f-4c07-fab7-5054ea00cdaa"
      },
      "source": [
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "clf = KNeighborsClassifier(n_neighbors=5) \n",
        "print(datetime.now().strftime(\"%H:%M:%S.%f\"))\n",
        "print(\"fitting...\")\n",
        "print(datetime.now().strftime(\"%H:%M:%S.%f\"))\n",
        "clf.fit(X_train,y_train)\n",
        "print(\"predicting training set...\")\n",
        "print(datetime.now().strftime(\"%H:%M:%S.%f\"))\n",
        "y_train_pred = clf.predict(X_train)\n",
        "print(\"predicting validation set...\")\n",
        "print(datetime.now().strftime(\"%H:%M:%S.%f\"))\n",
        "y_valid_pred = clf.predict(X_valid)\n",
        "print(\"predicting testing set...\")\n",
        "print(datetime.now().strftime(\"%H:%M:%S.%f\"))\n",
        "y_test_pred = clf.predict(X_test)\n",
        "print(\"training accuracy: {}\".format(accuracy_score(y_train, y_train_pred)))\n",
        "print(\"validation accuracy: {}\".format(accuracy_score(y_valid,y_valid_pred)))\n",
        "print(\"testing accuracy: {}\".format(accuracy_score(y_test,y_test_pred)))\n",
        "print(\"training accuracy (balanced): {}\".format(balanced_accuracy_score(y_train, y_train_pred)))\n",
        "print(\"validation accuracy (balanced): {}\".format(balanced_accuracy_score(y_valid, y_valid_pred)))\n",
        "print(\"testing accuracy (balanced): {}\".format(balanced_accuracy_score(y_test, y_test_pred)))\n",
        "print(datetime.now().strftime(\"%H:%M:%S.%f\"))\n",
        "#Confusion matrix whose i-th row and j-th column entry \n",
        "#indicates the number of samples with true label \n",
        "#being i-th class and predicted label being j-th class.\n",
        "cf_matrix_train = confusion_matrix(y_train_pred, y_train)\n",
        "print(\"confusion matrix for training set:\\n {}\".format(cf_matrix_train))\n",
        "sns.heatmap(cf_matrix_train, annot=False)\n",
        "plt.show()\n",
        "cf_matrix_valid = confusion_matrix(y_valid_pred, y_valid)\n",
        "print(\"confusion matrix for validation set:\\n {}\".format(cf_matrix_valid))\n",
        "sns.heatmap(cf_matrix_valid, annot=False,cmap=\"YlGnBu\")\n",
        "plt.show()\n",
        "\n",
        " "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "01:47:46.122867\n",
            "fitting...\n",
            "01:47:46.123899\n",
            "predicting training set...\n",
            "01:58:10.783731\n",
            "predicting validation set...\n",
            "05:14:05.478019\n",
            "predicting testing set...\n",
            "05:37:18.039911\n",
            "training accuracy: 0.9868759915440728\n",
            "validation accuracy: 0.9203868247240808\n",
            "testing accuracy: 0.9215424730858928\n",
            "training accuracy (balanced): 0.9868759915440728\n",
            "validation accuracy (balanced): 0.9055019767207738\n",
            "testing accuracy (balanced): 0.9078861036647291\n",
            "06:06:19.463122\n",
            "confusion matrix for training set:\n",
            " [[175103   6690      0      0      0      0      0]\n",
            " [  4989 171144      0      0      0      0      0]\n",
            " [    17    885 181419      0      0      0      0]\n",
            " [     0      4      7 181445      0      0      0]\n",
            " [   214   1678      0      0 181445      0      0]\n",
            " [    57    814     19      0      0 181445      0]\n",
            " [  1065    230      0      0      0      0 181445]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "confusion matrix for validation set:\n",
            " [[31316  2650     3     0    17     0    73]\n",
            " [ 2009 41431    34     0   104    38    12]\n",
            " [    6   312  5392    59    20   266     0]\n",
            " [    0     0    48   374     0    18     0]\n",
            " [   83   600    18     0  1350     6     0]\n",
            " [   15   294   277    30    10  2440     0]\n",
            " [  330    69     0     0     0     0  3258]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "confusion matrix for validation set:\n",
            " [[39511  3308     3     0    19     5    68]\n",
            " [ 2447 51612    33     0   125    49    12]\n",
            " [    5   342  6691    74    22   294     0]\n",
            " [    0     0    52   417     0    28     0]\n",
            " [  116   768    29     0  1820    12     1]\n",
            " [   20   372   313    35     9  3101     0]\n",
            " [  458    98     0     0     0     0  3934]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-9-061135a2414c>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     34\u001b[0m \u001b[0mcf_matrix_test\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mconfusion_matrix\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test_pred\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     35\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"confusion matrix for validation set:\\n {}\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcf_matrix_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 36\u001b[0;31m \u001b[0msns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mheatmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcf_test_valid\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mannot\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     37\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     38\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'cf_test_valid' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 401
        },
        "id": "hkDqoGDhDB4e",
        "outputId": "5ce06c40-59d0-4701-ec8a-af0242796448"
      },
      "source": [
        "cf_matrix_test = confusion_matrix(y_test_pred, y_test)\n",
        "print(\"confusion matrix for validation set:\\n {}\".format(cf_matrix_test))\n",
        "sns.heatmap(cf_matrix_test, annot=False)\n",
        "plt.show()\n",
        " "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "confusion matrix for validation set:\n",
            " [[39511  3308     3     0    19     5    68]\n",
            " [ 2447 51612    33     0   125    49    12]\n",
            " [    5   342  6691    74    22   294     0]\n",
            " [    0     0    52   417     0    28     0]\n",
            " [  116   768    29     0  1820    12     1]\n",
            " [   20   372   313    35     9  3101     0]\n",
            " [  458    98     0     0     0     0  3934]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}