Main file of Kernel K-Means implementation for DataMining - Clustering class in Coursera with my changes added.

main_kernel_k_means.py

import sys
from LoadData import * 
from k_means import * 
from evaluation import * 
from kernel_k_means import * 
import matplotlib
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D

if __name__ == "__main__":
    if len(sys.argv) != 4:
        print "[usage] <data-file> <ground-truth-file> <K-value>"
        exit(1) 
    

    dataFilename = sys.argv[1]
    groundtruthFilename = sys.argv[2]
    K = int(sys.argv[3])

    old_data = loadPoints(dataFilename) 
    groundtruth = loadClusters(groundtruthFilename) 

    sigma = 4.0

    #dimension of data
    nDimData = len(old_data[0])
    
    #number of data points
    nData = len(old_data) 

    data = kernel(old_data, sigma) 
    
    #size of groundtruth labels, R = 4 for [0,1,2,3] 
    R = len(set(groundtruth))
    
    print "size of data = %d  \ndimension of data = %d" % (nData,nDimData)
    
    
    #normalize groundtruth's classes to 0-indexed ones. [1,2,3] -> [0,1,2]
    min_val = min(set(groundtruth))
    if min_val>0:
        groundtruth = [x-min_val for x in groundtruth]

      

    centers = [] 
    for i in range(K):
        centers.append(data[i])
    
    results = kmeans(data, centers) 

    res_Purity = purity(results, groundtruth) 
    res_NMI = NMI(results, groundtruth) 
    
    
    if R>2:
        fig = plt.figure()
        ax = fig.add_subplot(111,projection='3d')
    colorarr=["b","r","y","g","m"]
    for k in range(nData):
        plt.hold(True)
        if R>2:
            ax.scatter(old_data[k][0], old_data[k][1],old_data[k][2],c=colorarr[results[k]])
        else:
            plt.scatter(old_data[k][0], old_data[k][1],c=colorarr[results[k]])

    #configurations to place the text top left of the plot
    x_axes = plt.gca().get_xlim()
    y_axes = plt.gca().get_ylim()

      
    if R>2:
        ax.text(x_axes[0] + 0.03*abs(x_axes[0]-x_axes[1]),y_axes[0] + 0.85*abs(y_axes[0]-y_axes[1]),0.9,
            "Method = %s\nFile = %s \nsize of Data = %dx%d \nK = %d \nsize of Ground Truth R = %d" %("Kernel K-Means",dataFilename,nData,nDimData,K,R),color="g")
    else:
        plt.text(x_axes[0] + 0.03*abs(x_axes[0]-x_axes[1]),y_axes[0] + 0.85*abs(y_axes[0]-y_axes[1]),
            "Method = %s\nFile = %s \nsize of Data = %dx%d \nK = %d \nsize of Ground Truth R = %d" %("Kernel K-Means",dataFilename,nData,nDimData,K,R),color="g")

    plt.show()
    #picname=dataFilename+"_Kernel_K-Means_" + str(K) + "_" + ".png"
    #matplotlib.pyplot.savefig(picname)
    print "Purity =", res_Purity
    print "NMI = ", res_NMI