Last active
June 9, 2018 15:05
-
-
Save grohith327/05f4e6d6ef6d24a76cbb8f5e38a8acd3 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
## K-Means Algorithm | |
import random | |
import numpy as np | |
## Randomly place the centroids of the three clusters | |
c1 = [float(np.random.randint(4,8)),float(np.random.randint(1,5)), | |
float(np.random.randint(1,7)),float(np.random.randint(0,3))] | |
c2 = [float(np.random.randint(4,8)),float(np.random.randint(1,5)), | |
float(np.random.randint(1,7)),float(np.random.randint(0,3))] | |
c3 = [float(np.random.randint(4,8)),float(np.random.randint(1,5)), | |
float(np.random.randint(1,7)),float(np.random.randint(0,3))] | |
## Intialize the number of iterations you want to run | |
epochs = 1 | |
while(epochs <= 100): | |
cluster_1 = [] | |
cluster_2 = [] | |
cluster_3 = [] | |
for point in train_data: | |
## Find the eucledian distance between all points the centroid | |
dis_point_c1 = ((c1[0]-point[0])**2 + (c1[1]-point[1])**2 + | |
(c1[2]-point[2])**2 + (c1[3]-point[3])**2)**0.5 | |
dis_point_c2 = ((c2[0]-point[0])**2 + (c2[1]-point[1])**2 + | |
(c2[2]-point[2])**2 + (c2[3]-point[3])**2)**0.5 | |
dis_point_c3 = ((c3[0]-point[0])**2 + (c3[1]-point[1])**2 + | |
(c3[2]-point[2])**2 + (c3[3]-point[3])**2)**0.5 | |
distances = [dis_point_c1,dis_point_c2,dis_point_c3] | |
## Find the closest centroid to the point and assign the point to that cluster | |
pos = distances.index(min(distances)) | |
if(pos == 0): | |
cluster_1.append(point) | |
elif(pos == 1): | |
cluster_2.append(point) | |
else: | |
cluster_3.append(point) | |
## Store the centroid values to calculate new centroid values | |
prev_c1 = c1 | |
prev_c2 = c2 | |
prev_c3 = c3 | |
cluster_1 = np.array(cluster_1) | |
cluster_2 = np.array(cluster_2) | |
cluster_3 = np.array(cluster_3) | |
## Find mean of all points within a cluster and make it as the centroid | |
if(len(cluster_1) != 0): | |
c1 = [sum(cluster_1[:,0])/float(len(cluster_1)), | |
sum(cluster_1[:,1])/float(len(cluster_1)), | |
sum(cluster_1[:,2])/float(len(cluster_1)), | |
sum(cluster_1[:,3])/float(len(cluster_1))] | |
if(len(cluster_2) != 0): | |
c2 = [sum(cluster_2[:,0])/float(len(cluster_2)), | |
sum(cluster_2[:,1])/float(len(cluster_2)), | |
sum(cluster_2[:,2])/float(len(cluster_2)), | |
sum(cluster_2[:,3])/float(len(cluster_2))] | |
if(len(cluster_3) != 0): | |
c3 = [sum(cluster_3[:,0])/float(len(cluster_3)), | |
sum(cluster_3[:,1])/float(len(cluster_3)), | |
sum(cluster_3[:,2])/float(len(cluster_3)), | |
sum(cluster_3[:,3])/float(len(cluster_3))] | |
## If centroid values hasn't changed, algorithm has convereged | |
if(prev_c1 == c1 and prev_c2 == c2 and prev_c3 == c3): | |
print("Converged") | |
break | |
print(epochs) | |
epochs += 1 | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment