selector.py

import GWO as gwo
import csv
import numpy
import time
from sklearn.model_selection import train_test_split
import pandas as pd
import fitnessFUNs
from sklearn.metrics import accuracy_score
from sklearn import svm



def selector(algo,func_details,popSize,Iter,completeData):
    function_name=func_details[0]
    lb=func_details[1]
    ub=func_details[2]
   
    
    DatasetSplitRatio=0.3  #Training 70%, Testing 30%
    
    DataFile=completeData
      
    data_set=numpy.loadtxt(open(DataFile,"rb"),delimiter=",",skiprows=0)
    numRowsData=numpy.shape(data_set)[0]    # number of instances in the  dataset
    numFeaturesData=numpy.shape(data_set)[1]-1 #number of features in the  dataset

    dataInput=data_set[0:numRowsData,0:-1] # 1: for class in first column, 0:-1 for class in last column
    dataTarget=data_set[0:numRowsData,-1]  # -1 for class in last column, 0 for class in first column
    trainInput, testInput, trainOutput, testOutput = train_test_split(dataInput, dataTarget, test_size=DatasetSplitRatio, random_state=1) 
    ratio=0.5
    validationInput, testInput, validationOutput, testOutput = train_test_split(testInput, testOutput, test_size=ratio, random_state=1)

    dim=numFeaturesData
    
    x=gwo.GWO(getattr(fitnessFUNs, function_name),lb,ub,dim,popSize,Iter,trainInput,trainOutput)
            
    reducedfeatures=[]
    for index in range(0,dim):
        if (x.bestIndividual[index]==1):
            reducedfeatures.append(index)
    reduced_data_train_global=trainInput[:,reducedfeatures]
    reduced_data_test_global=testInput[:,reducedfeatures]
    reduced_data_validation_global=validationInput[:,reducedfeatures]
    
    svc=svm.SVC(kernel='rbf').fit(reduced_data_train_global,trainOutput)

         # Compute the accuracy of the prediction
         
    target_pred_train = svc.predict(reduced_data_train_global)
    acc_train = float(accuracy_score(trainOutput, target_pred_train))
    x.trainAcc=acc_train
    
    target_pred_test = svc.predict(reduced_data_test_global)
    acc_test = float(accuracy_score(testOutput, target_pred_test))
    x.testAcc=acc_test
    
    target_pred_validation = svc.predict(reduced_data_validation_global)
    acc_val = float(accuracy_score(validationOutput, target_pred_validation))
    x.valAcc=acc_val
    
    
    return x
    
#####################################################################