Austin Dataset Gender Machine Learning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''numpy'''\n",
    "import numpy as np\n",
    "\n",
    "'''pandas'''\n",
    "import pandas as pd \n",
    "\n",
    "'''time'''\n",
    "import time\n",
    "\n",
    "'''seaborn'''\n",
    "import seaborn as sn\n",
    "\n",
    "'''sklearn'''\n",
    "from sklearn import neighbors, metrics \n",
    "from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV, RandomizedSearchCV, RepeatedStratifiedKFold\n",
    "from sklearn.preprocessing import LabelEncoder, StandardScaler\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.metrics import classification_report, confusion_matrix\n",
    "from sklearn.impute import SimpleImputer\n",
    "from sklearn.svm import SVC\n",
    "\n",
    "'''matplotlib'''\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "'''Remove warnings'''\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      PRIMARY_KEY    REP_DATE  REP_TIME SEX       APD_RACE_DESC  \\\n",
      "0       201612402  01/01/2016      2355   F               WHITE   \n",
      "1       201620049  01/02/2016       123   M  HISPANIC OR LATINO   \n",
      "2       201620049  01/02/2016       123   M  HISPANIC OR LATINO   \n",
      "3       201620049  01/02/2016       123   F  HISPANIC OR LATINO   \n",
      "4       201612438  01/01/2016      2212   M               WHITE   \n",
      "...           ...         ...       ...  ..                 ...   \n",
      "9179  20163651685  12/30/2016      2119   F               WHITE   \n",
      "9180  20163641610  12/29/2016      2320   M  HISPANIC OR LATINO   \n",
      "9181  20163660456  12/31/2016       817   M               WHITE   \n",
      "9182  20163651875  12/30/2016      2354   M  HISPANIC OR LATINO   \n",
      "9183  20163660189  12/31/2016       211   F               BLACK   \n",
      "\n",
      "                               LOCATION PERSON_SEARCHED_DESC  \\\n",
      "0               BURNET RD / W BRAKER LN              YES = 1   \n",
      "1     W STASSNEY LN / EMERALD FOREST DR              YES = 1   \n",
      "2     W STASSNEY LN / EMERALD FOREST DR               NO = 2   \n",
      "3     W STASSNEY LN / EMERALD FOREST DR              YES = 1   \n",
      "4       200 BLOCK W ANDERSON LN SVRD EB               NO = 2   \n",
      "...                                 ...                  ...   \n",
      "9179                1906 W SLAUGHTER LN              YES = 1   \n",
      "9180                3900 S CONGRESS AVE              YES = 1   \n",
      "9181                  2100 S LAMAR BLVD              YES = 1   \n",
      "9182               2501 S IH 35 SVRD NB              YES = 1   \n",
      "9183            100 E BEN WHITE BLVD EB               NO = 2   \n",
      "\n",
      "                               REASON_FOR_STOP_DESC  SEARCH_BASED_ON_DESC  \\\n",
      "0                                  CALL FOR SERVICE  INCIDENTAL TO ARREST   \n",
      "1                                  CALL FOR SERVICE        PROBABLE CAUSE   \n",
      "2                                               NaN                   NaN   \n",
      "3                           VIOLATION OF PENAL CODE        PROBABLE CAUSE   \n",
      "4                                               NaN                   NaN   \n",
      "...                                             ...                   ...   \n",
      "9179  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS  INCIDENTAL TO ARREST   \n",
      "9180  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS        PROBABLE CAUSE   \n",
      "9181  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS  INCIDENTAL TO ARREST   \n",
      "9182  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS  INCIDENTAL TO ARREST   \n",
      "9183  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS                   NaN   \n",
      "\n",
      "     SEARCH_DISC_DESC                                        RACE_KNOWN  \\\n",
      "0             NOTHING  NO - RACE OR ETHNICITY WAS NOT KNOWN BEFORE STOP   \n",
      "1               OTHER  NO - RACE OR ETHNICITY WAS NOT KNOWN BEFORE STOP   \n",
      "2                 NaN                                               NaN   \n",
      "3             NOTHING     YES - RACE OR ETHNICITY WAS KNOWN BEFORE STOP   \n",
      "4                 NaN                                               NaN   \n",
      "...               ...                                               ...   \n",
      "9179          NOTHING  NO - RACE OR ETHNICITY WAS NOT KNOWN BEFORE STOP   \n",
      "9180            DRUGS  NO - RACE OR ETHNICITY WAS NOT KNOWN BEFORE STOP   \n",
      "9181          NOTHING  NO - RACE OR ETHNICITY WAS NOT KNOWN BEFORE STOP   \n",
      "9182          NOTHING  NO - RACE OR ETHNICITY WAS NOT KNOWN BEFORE STOP   \n",
      "9183              NaN  NO - RACE OR ETHNICITY WAS NOT KNOWN BEFORE STOP   \n",
      "\n",
      "      X_COORDINATE  Y_COORDINATE   SECTOR  LOCAL_FIELD1  \n",
      "0          3119888      10115666  ADAM PD           7.0  \n",
      "1          3100814      10049567    FRANK           2.0  \n",
      "2          3100814      10049567    FRANK           2.0  \n",
      "3          3100814      10049567    FRANK           2.0  \n",
      "4          3125331      10098656      IDA           4.0  \n",
      "...            ...           ...      ...           ...  \n",
      "9179       3089803      10035918    FRANK           5.0  \n",
      "9180       3108121      10054692    DAVID           3.0  \n",
      "9181       3105636      10063425    DAVID           5.0  \n",
      "9182       3115068      10057440     HENR           3.0  \n",
      "9183       3108116      10054117    DAVID           3.0  \n",
      "\n",
      "[9184 rows x 15 columns]\n"
     ]
    }
   ],
   "source": [
    "'''Read and display data'''\n",
    "data = pd.read_csv(\"Datasets/2016-rp-arrests-1.csv\") #Load in the arrest data\n",
    "print(data) #Display the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "     SEX       APD_RACE_DESC PERSON_SEARCHED_DESC  \\\n",
      "0      F               WHITE              YES = 1   \n",
      "1      M  HISPANIC OR LATINO              YES = 1   \n",
      "2      M  HISPANIC OR LATINO               NO = 2   \n",
      "3      F  HISPANIC OR LATINO              YES = 1   \n",
      "4      M               WHITE               NO = 2   \n",
      "...   ..                 ...                  ...   \n",
      "9179   F               WHITE              YES = 1   \n",
      "9180   M  HISPANIC OR LATINO              YES = 1   \n",
      "9181   M               WHITE              YES = 1   \n",
      "9182   M  HISPANIC OR LATINO              YES = 1   \n",
      "9183   F               BLACK               NO = 2   \n",
      "\n",
      "                               REASON_FOR_STOP_DESC  SEARCH_BASED_ON_DESC  \n",
      "0                                  CALL FOR SERVICE  INCIDENTAL TO ARREST  \n",
      "1                                  CALL FOR SERVICE        PROBABLE CAUSE  \n",
      "2                                               NaN                   NaN  \n",
      "3                           VIOLATION OF PENAL CODE        PROBABLE CAUSE  \n",
      "4                                               NaN                   NaN  \n",
      "...                                             ...                   ...  \n",
      "9179  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS  INCIDENTAL TO ARREST  \n",
      "9180  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS        PROBABLE CAUSE  \n",
      "9181  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS  INCIDENTAL TO ARREST  \n",
      "9182  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS  INCIDENTAL TO ARREST  \n",
      "9183  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS                   NaN  \n",
      "\n",
      "[9184 rows x 5 columns]\n"
     ]
    }
   ],
   "source": [
    "'''Assigning attributes to use'''\n",
    "#data = data.head(n=100)\n",
    "\n",
    "dataColumns = [ #Select all columns needed \n",
    "    'SEX',\n",
    "    'APD_RACE_DESC',\n",
    "    'PERSON_SEARCHED_DESC',\n",
    "    'REASON_FOR_STOP_DESC',\n",
    "    'SEARCH_BASED_ON_DESC']\n",
    "\n",
    "selectedClass = ['SEX'] #Assign the class column\n",
    "\n",
    "selectedData = data[dataColumns].values \n",
    "\n",
    "selectedData = pd.DataFrame(selectedData, columns = [dataColumns]) #Make into pandas dataframe\n",
    "print(selectedData)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "     SEX       APD_RACE_DESC PERSON_SEARCHED_DESC  \\\n",
      "0      F               WHITE                 TRUE   \n",
      "1      M  HISPANIC OR LATINO                 TRUE   \n",
      "2      M  HISPANIC OR LATINO                FALSE   \n",
      "3      F  HISPANIC OR LATINO                 TRUE   \n",
      "4      M               WHITE                FALSE   \n",
      "...   ..                 ...                  ...   \n",
      "9179   F               WHITE                 TRUE   \n",
      "9180   M  HISPANIC OR LATINO                 TRUE   \n",
      "9181   M               WHITE                 TRUE   \n",
      "9182   M  HISPANIC OR LATINO                 TRUE   \n",
      "9183   F               BLACK                FALSE   \n",
      "\n",
      "                               REASON_FOR_STOP_DESC  SEARCH_BASED_ON_DESC  \n",
      "0                                  CALL FOR SERVICE  INCIDENTAL TO ARREST  \n",
      "1                                  CALL FOR SERVICE        PROBABLE CAUSE  \n",
      "2                                    NONE SPECIFIED        NONE SPECIFIED  \n",
      "3                           VIOLATION OF PENAL CODE        PROBABLE CAUSE  \n",
      "4                                    NONE SPECIFIED        NONE SPECIFIED  \n",
      "...                                             ...                   ...  \n",
      "9179  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS  INCIDENTAL TO ARREST  \n",
      "9180  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS        PROBABLE CAUSE  \n",
      "9181  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS  INCIDENTAL TO ARREST  \n",
      "9182  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS  INCIDENTAL TO ARREST  \n",
      "9183  VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS        NONE SPECIFIED  \n",
      "\n",
      "[9184 rows x 5 columns]\n"
     ]
    }
   ],
   "source": [
    "#selectedData[['REASON_FOR_STOP_DESC']] = selectedData[['REASON_FOR_STOP_DESC']].fillna(value=\"NONE SPECIFIED\")\n",
    "\n",
    "selectedData = selectedData.replace(\"YES = 1\", \"TRUE\") #Making the data a bit cleaner\n",
    "selectedData = selectedData.replace(\"NO = 2\", \"FALSE\")\n",
    "selectedData.fillna(\"NONE SPECIFIED\", inplace=True) #I think that the absense might be useful so I'm keeping it in for this\n",
    "\n",
    "print(selectedData)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SEX\n"
     ]
    }
   ],
   "source": [
    "print(selectedClass[0]) #Checking the right class is selected after"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[['WHITE' 'TRUE' 'CALL FOR SERVICE' 'INCIDENTAL TO ARREST']\n",
      " ['HISPANIC OR LATINO' 'TRUE' 'CALL FOR SERVICE' 'PROBABLE CAUSE']\n",
      " ['HISPANIC OR LATINO' 'FALSE' 'NONE SPECIFIED' 'NONE SPECIFIED']\n",
      " ...\n",
      " ['WHITE' 'TRUE' 'VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS'\n",
      "  'INCIDENTAL TO ARREST']\n",
      " ['HISPANIC OR LATINO' 'TRUE'\n",
      "  'VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS' 'INCIDENTAL TO ARREST']\n",
      " ['BLACK' 'FALSE' 'VIOLATION OF TRANSPORTATION CODE/VEHICLE LAWS'\n",
      "  'NONE SPECIFIED']]      SEX\n",
      "0      F\n",
      "1      M\n",
      "2      M\n",
      "3      F\n",
      "4      M\n",
      "...   ..\n",
      "9179   F\n",
      "9180   M\n",
      "9181   M\n",
      "9182   M\n",
      "9183   F\n",
      "\n",
      "[9184 rows x 1 columns]\n",
      "(9184, 4) (9184, 1)\n",
      "{'M', 'F'}\n"
     ]
    }
   ],
   "source": [
    "dataColumns.remove(selectedClass[0]) #Remove the class from the columns needed for training\n",
    "\n",
    "X = selectedData[dataColumns].values \n",
    "\n",
    "#Set the class\n",
    "y = selectedData[selectedClass]\n",
    "\n",
    "#X = selectedData.to_numpy()\n",
    "print(X, y)\n",
    "print(X.shape, y.shape)\n",
    "\n",
    "realYValues = y\n",
    "print(set(np.ravel(realYValues)))\n",
    "\n",
    "y = np.ravel(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'M', 'F'}\n",
      "[[7 1 0 3]\n",
      " [4 1 0 6]\n",
      " [4 0 3 5]\n",
      " ...\n",
      " [7 1 9 3]\n",
      " [4 1 9 3]\n",
      " [2 0 9 5]] [0 1 1 ... 1 1 0]\n"
     ]
    }
   ],
   "source": [
    "'''Translate classification values into numerical values'''\n",
    "Le = LabelEncoder() #Use the LabelEncoder library\n",
    "for i in range(len(X[0])): #Iterate over instances of the data\n",
    "    X[:, i] = Le.fit_transform(X[:, i]) #Use fit_transform to convert values \n",
    "\n",
    "print(set(y))\n",
    "\n",
    "classificationValues = set(y)\n",
    "\n",
    "yRealValues, y = np.unique(y, return_inverse=True)\n",
    "\n",
    "y = np.ravel(y) #Put the classification into a single 1D array to avoid future warning and error messages\n",
    "print(X, y) #Ensure that the transformations have taken place correctly"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''Simple function to train the models from the data'''\n",
    "def trainModel(model):\n",
    "    model.fit(X_train, y_train) #Fit the model\n",
    "    prediction = model.predict(X_test) #Give the predictions for the y values\n",
    "    return round(metrics.accuracy_score(y_test, prediction), 3), classification_report(y_test, prediction) #Return the accuracy value and the report for the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "knn = neighbors.KNeighborsClassifier(n_neighbors=17, weights='uniform') #Define the KNN algorithm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2) #Split the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def trainData(model):\n",
    "    start_time = time.time()  #Set the starting execution time\n",
    "    acc, rep = trainModel(model) #set knn accuracy and report values on uncleaned data\n",
    "    finish_time = round(time.time() - start_time, 3)\n",
    "    print(\"{} seconds to run for {}\".format(finish_time, model)) #Display the runtime\n",
    "    return acc, rep"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''Just run all the models in order to get results for all of them'''\n",
    "def runAllModels():\n",
    "    knn_acc, knn_rep = trainData(knn)\n",
    "    log_acc, log_rep = trainData(LogisticRegression())\n",
    "    svc_acc, svc_rep = trainData(SVC())\n",
    "    return knn_acc, knn_rep, log_acc, log_rep, svc_acc, svc_rep"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def displayAllModels():\n",
    "    knn_acc, knn_rep, log_acc, log_rep, svc_acc, svc_rep = runAllModels()\n",
    "    print(\"\\nKNN\\n--------\\nAccuracy = {}\\n{}\".format(knn_acc, knn_rep)) #Display the accuracy and report for the entire dataset for knn\n",
    "    print(\"Logistic Regression\\n--------\\nAccuracy = {}\\n{}\".format(log_acc, log_rep)) #Display the accuracy and report for the entire dataset for logistic regression\n",
    "    print(\"SVC\\n--------\\nAccuracy = {}\\n{}\".format(svc_acc, svc_rep)) #Display the accuracy and report for the entire dataset for SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.117 seconds to run for KNeighborsClassifier(n_neighbors=17)\n",
      "0.037 seconds to run for LogisticRegression()\n",
      "1.293 seconds to run for SVC()\n",
      "\n",
      "KNN\n",
      "--------\n",
      "Accuracy = 0.739\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.32      0.08      0.13       438\n",
      "           1       0.77      0.94      0.85      1399\n",
      "\n",
      "    accuracy                           0.74      1837\n",
      "   macro avg       0.54      0.51      0.49      1837\n",
      "weighted avg       0.66      0.74      0.68      1837\n",
      "\n",
      "Logistic Regression\n",
      "--------\n",
      "Accuracy = 0.762\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00       438\n",
      "           1       0.76      1.00      0.86      1399\n",
      "\n",
      "    accuracy                           0.76      1837\n",
      "   macro avg       0.38      0.50      0.43      1837\n",
      "weighted avg       0.58      0.76      0.66      1837\n",
      "\n",
      "SVC\n",
      "--------\n",
      "Accuracy = 0.762\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00       438\n",
      "           1       0.76      1.00      0.86      1399\n",
      "\n",
      "    accuracy                           0.76      1837\n",
      "   macro avg       0.38      0.50      0.43      1837\n",
      "weighted avg       0.58      0.76      0.66      1837\n",
      "\n"
     ]
    }
   ],
   "source": [
    "displayAllModels()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = StandardScaler() #Define which scaler to use\n",
    "scale_X = scaler.fit_transform(X) #Scale the entire dataset\n",
    "X_train, X_test, y_train, y_test = train_test_split(scale_X, y, test_size=0.2) #Split the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.123 seconds to run for KNeighborsClassifier(n_neighbors=17)\n",
      "0.011 seconds to run for LogisticRegression()\n",
      "1.272 seconds to run for SVC()\n",
      "\n",
      "KNN\n",
      "--------\n",
      "Accuracy = 0.723\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.34      0.08      0.12       476\n",
      "           1       0.75      0.95      0.84      1361\n",
      "\n",
      "    accuracy                           0.72      1837\n",
      "   macro avg       0.54      0.51      0.48      1837\n",
      "weighted avg       0.64      0.72      0.65      1837\n",
      "\n",
      "Logistic Regression\n",
      "--------\n",
      "Accuracy = 0.741\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00       476\n",
      "           1       0.74      1.00      0.85      1361\n",
      "\n",
      "    accuracy                           0.74      1837\n",
      "   macro avg       0.37      0.50      0.43      1837\n",
      "weighted avg       0.55      0.74      0.63      1837\n",
      "\n",
      "SVC\n",
      "--------\n",
      "Accuracy = 0.741\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00       476\n",
      "           1       0.74      1.00      0.85      1361\n",
      "\n",
      "    accuracy                           0.74      1837\n",
      "   macro avg       0.37      0.50      0.43      1837\n",
      "weighted avg       0.55      0.74      0.63      1837\n",
      "\n"
     ]
    }
   ],
   "source": [
    "displayAllModels() #Scaled data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cross_val(model):\n",
    "    accuracies = cross_val_score(model, scale_X, y, cv=5, scoring='accuracy') #Cross validate the machine learning model\n",
    "    return \"Cross validated average scaled accuracy = {}\\nMaximum scaled accuracy = {}\\nMinimum scaled accuracy = {}\\n\".format(round(accuracies.mean(), 3), round(accuracies.max(), 3), round(accuracies.min(), 3)) #Return the maximum and average accuracy using cross validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def displayCrossValidation():\n",
    "    print(\"KNN\")\n",
    "    print(cross_val(knn)) #Display the cross validation score for knn using the entire dataset\n",
    "    print(\"\\nLogistic Regression\")\n",
    "    print(cross_val(LogisticRegression())) #Display the cross validation score for logistic regression using the entire dataset\n",
    "    print(\"\\nSVC\")\n",
    "    print(cross_val(SVC())) #Display the cross validation score for logistic regression using the entire dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KNN\n",
      "Cross validated average scaled accuracy = 0.734\n",
      "Maximum scaled accuracy = 0.751\n",
      "Minimum scaled accuracy = 0.675\n",
      "\n",
      "\n",
      "Logistic Regression\n",
      "Cross validated average scaled accuracy = 0.75\n",
      "Maximum scaled accuracy = 0.75\n",
      "Minimum scaled accuracy = 0.75\n",
      "\n",
      "\n",
      "SVC\n",
      "Cross validated average scaled accuracy = 0.75\n",
      "Maximum scaled accuracy = 0.75\n",
      "Minimum scaled accuracy = 0.75\n",
      "\n"
     ]
    }
   ],
   "source": [
    "displayCrossValidation()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''Set the parameters for knn'''\n",
    "n_neighbors = range(1, 21, 2) #Sets a range for the number of neighbours to check against\n",
    "weights = ['uniform', 'distance'] #Checks the difference between distance and uniform usage of knn\n",
    "metric = ['euclidean', 'manhattan', 'minkowski'] #Set the knn metrics to check\n",
    "\n",
    "'''Set variables for grid search'''\n",
    "solvers = ['newton-cg', 'lbfgs', 'liblinear'] #Sets the solvers used to optimise the algrorithm\n",
    "penalty = ['l1', 'l2'] #Used to penalise the hyperparamter\n",
    "c_values = [100, 10, 1.0, 0.1, 0.01, 0.001] #Sets the strength of the penalty\n",
    "\n",
    "cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=3, random_state=1) #Set the cross validation parameters\n",
    "\n",
    "knn_grid = dict(n_neighbors=n_neighbors,weights=weights,metric=metric) #Set the parameters for knn\n",
    "log_grid = dict(solver=solvers, penalty=penalty, C=c_values) #Set the parameters for logistic regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''Functions to run grid search and random search'''\n",
    "def display_grid_search():\n",
    "    start_time = time.time() #Set the starting execution time\n",
    "    grid_search = GridSearchCV(estimator=model, param_grid=grid, n_jobs=-1, cv=cv, scoring='accuracy',error_score=0) #Define the grid search\n",
    "    grid_result = grid_search.fit(X, y) #Fit the model\n",
    "    print(\"{} seconds to run\".format(round(time.time() - start_time, 3))) #Display the runtime \n",
    "    print(\"Best accuracy {} using {}\".format(round(grid_result.best_score_, 3), grid_result.best_params_)) #Display the best score and the best parameters for grid search\n",
    "def display_random_search():\n",
    "    start_time = time.time()  #Set the starting execution time\n",
    "    rand_search = RandomizedSearchCV(model, grid, n_jobs=-1, cv=cv, scoring='accuracy',error_score=0) #Define the random search\n",
    "    rand_result = rand_search.fit(X, y) #Fit the model\n",
    "    print(\"{} seconds to run\".format(round(time.time() - start_time, 3))) #Display the runtime\n",
    "    print(\"Best accuracy {} using {}\".format(round(rand_result.best_score_, 3), rand_result.best_params_)) #Display the best score and the best parameters for random search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "35.744 seconds to run\n",
      "Best accuracy 0.747 using {'metric': 'manhattan', 'n_neighbors': 17, 'weights': 'uniform'}\n"
     ]
    }
   ],
   "source": [
    "'''Grid search with KNN'''\n",
    "grid = knn_grid #Set the grid as the ones defined for knn\n",
    "model = KNeighborsClassifier() #Set the model to knn\n",
    "display_grid_search()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5.722 seconds to run\n",
      "Best accuracy 0.745 using {'weights': 'uniform', 'n_neighbors': 15, 'metric': 'minkowski'}\n"
     ]
    }
   ],
   "source": [
    "'''Random search with KNN'''\n",
    "grid = knn_grid #Set the grid as the ones defined for knn\n",
    "model = KNeighborsClassifier() #Set the model to knn\n",
    "display_random_search()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8.813 seconds to run\n",
      "Best accuracy 0.75 using {'C': 100, 'penalty': 'l1', 'solver': 'liblinear'}\n"
     ]
    }
   ],
   "source": [
    "'''Grid search with Logistic Regression'''\n",
    "grid = log_grid #Set the grid as the ones defined for logisic regression\n",
    "model = LogisticRegression() #Set the model to logistic regression\n",
    "display_grid_search()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.22 seconds to run\n",
      "Best accuracy 0.75 using {'solver': 'newton-cg', 'penalty': 'l2', 'C': 0.001}\n"
     ]
    }
   ],
   "source": [
    "'''Random search with Logistic Regression'''\n",
    "grid = log_grid #Set the grid as the ones defined for logisic regression\n",
    "model = LogisticRegression() #Set the model to logistic regression\n",
    "display_random_search()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''Plot a confusion matrix'''\n",
    "def c_m():\n",
    "    start_time = time.time()  #Set the starting execution time\n",
    "    y_pred = accpred() #Get the predicted values for y\n",
    "    matrix = confusion_matrix(y_test, y_pred) #Build a confusion matrix for the predicted and actual values\n",
    "    real_vals = assignRealClasses()\n",
    "    df_cm = pd.DataFrame(matrix, columns=np.unique(real_vals), index = np.unique(real_vals)) #Set the columns as the real values\n",
    "    df_cm.index.name = 'Actual' #Label the x axis as Actual\n",
    "    df_cm.columns.name = 'Predicted' #Label the y axis as Predicted\n",
    "    plt.figure(figsize = (10,7)) #Set the size\n",
    "    sn.set(font_scale=1.4) #Label size\n",
    "    sn.heatmap(df_cm, cmap=\"Blues\", annot=True,annot_kws={\"size\": 16}) #Font size\n",
    "    finish_time = round(time.time() - start_time, 3)\n",
    "    print(\"{} seconds to run\".format(finish_time)) #Display the runtime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def accpred():\n",
    "    model.fit(X_train, y_train) #Fit the model\n",
    "    return model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "realClasses = {\n",
    "    0:'FEMALE', \n",
    "    1:'MALE'\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''Assign the real string data to the numerical values used for training data'''\n",
    "def assignRealClasses(): \n",
    "    real_vals = []\n",
    "    for i in set(y_test):\n",
    "        real_vals.append(realClasses[i])\n",
    "    return real_vals"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{0, 1}\n"
     ]
    }
   ],
   "source": [
    "print(set(y_test)) #Check to see which values have been tested"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KNN\n",
      "0.239 seconds to run\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "'''Display the confusion matrix for knn'''\n",
    "model = knn #Set the model as knn\n",
    "print(\"KNN\")\n",
    "c_m()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      "0.075 seconds to run\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "'''Display the confusion matrix for logistic regession'''\n",
    "model = LogisticRegression() #Set the model as logistic regression\n",
    "print(\"Logistic Regression\")\n",
    "c_m()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SVC\n",
      "1.259 seconds to run\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "'''Display the confusion matrix for logistic regession'''\n",
    "model = SVC() #Set the model as logistic regression\n",
    "print(\"SVC\")\n",
    "c_m()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}