KNN_archive.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "CLMCL3IAoIn2"
   },
   "source": [
    "##K nearest neighbour\n",
    " \n",
    "- run the k nearest neighbour algorithm for a range of values of k\n",
    "- output the prediction and compare with the true values on the validation set\n",
    "- choose best k and run the k nearest neighbour with smallest error on the validation set and use this k to compute estimate of accuracy on the test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "id": "Hmdd_d8eoALz"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "def data_cleaner(data):\n",
    "    #remove the id, partlybad columns\n",
    "    data.drop([\"id\",\"partlybad\"],1,inplace=True)\n",
    "    #set new index using the date column\n",
    "    data.set_index([\"date\"],inplace=True)\n",
    "    #add binary 'class2' col for event/non-event\n",
    "    data[\"class2\"] = np.where(data[\"class4\"] == \"nonevent\", \"nonevent\",\"event\")\n",
    "    #arrange to put class2 col at the front\n",
    "    cols = ['class2'] + [col for col in data if col != 'class2']\n",
    "    data = data[cols]\n",
    "    return data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "id": "9vBoplXwoLlz"
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "#load training data\n",
    "url = \"https://raw.githubusercontent.com/williwilliams3/TermProjectIML/master/data/train.csv\"\n",
    "npf_train_raw = pd.read_csv(url)\n",
    "npf_train = data_cleaner(npf_train_raw)\n",
    "\n",
    "#create feature array and target column for training data\n",
    "X_train = npf_train.iloc[:,2:]\n",
    "y_train = npf_train[\"class2\"]\n",
    "\n",
    "#import validation data\n",
    "npf_validation_raw = pd.read_csv(\"https://raw.githubusercontent.com/williwilliams3/TermProjectIML/master/data/validation.csv\")\n",
    "npf_validation = data_cleaner(npf_validation_raw)\n",
    "\n",
    "#create feature array and target column for test data\n",
    "X_validation = npf_validation.iloc[:,2:]\n",
    "y_validation = npf_validation[\"class2\"]\n",
    "\n",
    "#import test data\n",
    "npf_test_raw = pd.read_csv(\"https://raw.githubusercontent.com/williwilliams3/TermProjectIML/master/data/test.csv\")\n",
    "npf_test = data_cleaner(npf_test_raw)\n",
    "\n",
    "#create feature array and target column for test data\n",
    "X_test = npf_test.iloc[:,2:]\n",
    "y_test = npf_test[\"class2\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "pXRDOp8SoQSd"
   },
   "source": [
    "Run the k-nn algorithm for various values of k. Each iteration we are fitting the model to the training set, and then calculating the error on the validation set. Then we take the value of k with lowest error to fit the model to the test set. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 969
    },
    "id": "Y3If3t6ZoTDS",
    "outputId": "386f4b1d-ea67-4176-fdaa-5a722f959e82"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rate of correct predictions for k = 1 0.7441860465116279\n",
      "rate of correct predictions for k = 2 0.7441860465116279\n",
      "rate of correct predictions for k = 3 0.7558139534883721\n",
      "rate of correct predictions for k = 4 0.7209302325581395\n",
      "rate of correct predictions for k = 5 0.7441860465116279\n",
      "rate of correct predictions for k = 6 0.7325581395348837\n",
      "rate of correct predictions for k = 7 0.7558139534883721\n",
      "rate of correct predictions for k = 8 0.7790697674418605\n",
      "rate of correct predictions for k = 9 0.7441860465116279\n",
      "rate of correct predictions for k = 10 0.7558139534883721\n",
      "rate of correct predictions for k = 11 0.7441860465116279\n",
      "rate of correct predictions for k = 12 0.7325581395348837\n",
      "rate of correct predictions for k = 13 0.7093023255813954\n",
      "rate of correct predictions for k = 14 0.7325581395348837\n",
      "rate of correct predictions for k = 15 0.7441860465116279\n",
      "rate of correct predictions for k = 16 0.7558139534883721\n",
      "rate of correct predictions for k = 17 0.7558139534883721\n",
      "rate of correct predictions for k = 18 0.7674418604651163\n",
      "rate of correct predictions for k = 19 0.7558139534883721\n",
      "rate of correct predictions for k = 20 0.7674418604651163\n",
      "rate of correct predictions for k = 21 0.7558139534883721\n",
      "rate of correct predictions for k = 22 0.7674418604651163\n",
      "rate of correct predictions for k = 23 0.7674418604651163\n",
      "rate of correct predictions for k = 24 0.7674418604651163\n",
      "rate of correct predictions for k = 25 0.7674418604651163\n",
      "rate of correct predictions for k = 26 0.7674418604651163\n",
      "rate of correct predictions for k = 27 0.7558139534883721\n",
      "rate of correct predictions for k = 28 0.7558139534883721\n",
      "rate of correct predictions for k = 29 0.7558139534883721\n",
      "rate of correct predictions for k = 30 0.7441860465116279\n",
      "rate of correct predictions for k = 31 0.7441860465116279\n",
      "rate of correct predictions for k = 32 0.7558139534883721\n",
      "rate of correct predictions for k = 33 0.7441860465116279\n",
      "rate of correct predictions for k = 34 0.7558139534883721\n",
      "rate of correct predictions for k = 35 0.7441860465116279\n",
      "rate of correct predictions for k = 36 0.7325581395348837\n",
      "rate of correct predictions for k = 37 0.7209302325581395\n",
      "rate of correct predictions for k = 38 0.7209302325581395\n",
      "rate of correct predictions for k = 39 0.7209302325581395\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(1.0, 40.0)"
      ]
     },
     "execution_count": 17,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#k nearest neighbours\n",
    "from sklearn.neighbors import KNeighborsClassifier \n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "#set up for storing error rate\n",
    "pred_rate = np.zeros(40)\n",
    "\n",
    "for k in range(1,40):\n",
    "    #run algorithm\n",
    "    knn = KNeighborsClassifier(n_neighbors=k) \n",
    "    #fit to training data\n",
    "    knn.fit(X_train, y_train) \n",
    "    #predicted values for class2 on validation set\n",
    "    y_predictions = knn.predict(X_validation)\n",
    "    #compare predictions with actual results\n",
    "    y_s = np.c_[y_validation,y_predictions]\n",
    "    correct_ind = np.where(y_s[:,0] == y_s[:,1], 1,0)\n",
    "    #find error on validation set\n",
    "    pred_rate[k] = np.mean(correct_ind)\n",
    "    print('rate of correct predictions for k = %i' %k,pred_rate[k])\n",
    "\n",
    "f, ax = plt.subplots(1)\n",
    "plt.plot(pred_rate,color=\"fuchsia\")\n",
    "ax.set_xlim(1, 40)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 314
    },
    "id": "eQgoaJD-oWtO",
    "outputId": "6e6970a7-7e8c-4853-d919-09ba637ef9c9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "binary accuracy for 13-NN on test set= 0.8604651162790697\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7fa95887d588>"
      ]
     },
     "execution_count": 18,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "#use k with lowest validation error \n",
    "#k=13\n",
    "knn = KNeighborsClassifier(n_neighbors=13) \n",
    "#fit to training data\n",
    "knn.fit(X_train, y_train) \n",
    "\n",
    "y_predictions = knn.predict(X_test)\n",
    "\n",
    "#accuracy rate\n",
    "from sklearn import metrics\n",
    "#accuracy metric for the test set \n",
    "print('binary accuracy for 13-NN on test set=', metrics.accuracy_score(y_test, y_predictions))\n",
    "\n",
    "#map predictions on heat map\n",
    "confusion_matrix = pd.crosstab(y_test, y_predictions, rownames=['Actual'], colnames=['Predicted'])\n",
    "sns.heatmap(confusion_matrix, annot=True)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "N5XNyp7Yoe0W"
   },
   "source": [
    "Results\n",
    "- Lowest validation error was given by k=13\n",
    "- classifies well for non-event, less well for event days, predicts highly that they will be non-event. \n",
    "\n",
    "However, there is some inconsistency in the validation error and the test error. We would expect this to be similar, since the validation error should predict the error on the test set. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 297
    },
    "id": "nyAycM4GofCm",
    "outputId": "01dd94cd-532a-4b3b-ae96-32a191489dcc"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy for 13-NN on test set= 0.6627906976744186\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#repeat optimal k finding for multiclassifier\n",
    "#try multi classifiers with k nearest\n",
    "\n",
    "y_train_multi = npf_train[\"class4\"]\n",
    "\n",
    "y_test_multi = npf_test[\"class4\"]\n",
    "\n",
    "knn = KNeighborsClassifier(n_neighbors=13) \n",
    "knn.fit(X_train, y_train_multi) \n",
    "\n",
    "#predicted values for class4\n",
    "y_predictions_multi = knn.predict(X_test)\n",
    "\n",
    "confusion_matrix = pd.crosstab(y_test_multi, y_predictions_multi, rownames=['Actual'], colnames=['Predicted'])\n",
    "sns.heatmap(confusion_matrix, annot=True)\n",
    "\n",
    "\n",
    "print('accuracy for 13-NN on test set=', metrics.accuracy_score(y_test_multi, y_predictions_multi))"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "name": "KNN_archive.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}