Update Clustering.py

Clustering.py

@@ -7,28 +7,30 @@
 import numpy as np 
 import pandas as pd 
 import matplotlib.pyplot as plt 
+import matplotlib.cm as cm
 import sys 
 
 # creating data 
 mean_01 = np.array([0.0, 0.0]) 
 cov_01 = np.array([[1, 0.3], [0.3, 1]]) 
-dist_01 = np.random.multivariate_normal(mean_01, cov_01, 100) 
+dist_01 = np.random.multivariate_normal(mean_01, cov_01, 10) 
 
 mean_02 = np.array([6.0, 7.0]) 
 cov_02 = np.array([[1.5, 0.3], [0.3, 1]]) 
-dist_02 = np.random.multivariate_normal(mean_02, cov_02, 100) 
+dist_02 = np.random.multivariate_normal(mean_02, cov_02, 10) 
 
 mean_03 = np.array([7.0, -5.0]) 
 cov_03 = np.array([[1.2, 0.5], [0.5, 1,3]]) 
-dist_03 = np.random.multivariate_normal(mean_03, cov_01, 100) 
+dist_03 = np.random.multivariate_normal(mean_03, cov_01, 10) 
 
 mean_04 = np.array([2.0, -7.0]) 
 cov_04 = np.array([[1.2, 0.5], [0.5, 1,3]]) 
-dist_04 = np.random.multivariate_normal(mean_04, cov_01, 100) 
+dist_04 = np.random.multivariate_normal(mean_04, cov_01, 10) 
 
 data = np.vstack((dist_01, dist_02, dist_03, dist_04)) 
 np.random.shuffle(data) 
 
+
 
 # function to compute euclidean distance 
 def distance(p1, p2): 
@@ -46,30 +48,30 @@ def initialize(data, no_of_clusters):
     ## a randomly selected data point to the list 
     centroids = [] 
     centroids.append(data[np.random.randint( 
-            data.shape[0]), :]) 
+            data.shape[0]), :].tolist())
+    temp_data=data.tolist()
+    temp_data.append('dummy')
+
 
     ## compute remaining no_of_clusters - 1 centroids 
-    for c_id in range(no_of_clusters - 1): 
+    for c_id in range(no_of_clusters-1): 
 
         ## initialize a list to store distances of data 
         ## points from nearest centroid 
         dist = [] 
-        for i in range(data.shape[0]): 
-            point = data[i, :] 
-            d = sys.maxsize 
-
-            ## compute distance of 'point' from each of the previously 
-            ## selected centroid and store the minimum distance 
-            for j in range(len(centroids)): 
-                temp_dist = distance(point, centroids[j]) 
-                d = min(d, temp_dist) 
-            dist.append(d) 
-
-        ## select data point with maximum distance as our next centroid 
-        dist = np.array(dist) 
-        next_centroid = data[np.argmax(dist), :] 
-        centroids.append(next_centroid) 
-        dist = []  
+        for i in range(data.shape[0]):
+            if [data[i,:].tolist()]*len(centroids)!=centroids:
+                temp_sum=1
+                for j in range(c_id+1):
+                    #dist.append(distance(data[i,:],centroids[c_id]))
+                    temp_sum *= distance(data[i,:],centroids[j])
+                dist.append(temp_sum)
+            else:
+                dist.append(0)
+        pdf=(dist/sum(dist)).tolist()
+        pdf.append(0)
+        next_centroid = np.random.choice(temp_data,p=pdf)
+        centroids.append(next_centroid)
     return centroids 
 
 def classify_a_point(point, groups, k):
@@ -99,20 +101,41 @@ def classify_a_point(point, groups, k):
 
 def cluster(data, no_of_clusters, k, max_iterations):
     groups=initialize(data, no_of_clusters)
+    #print('groups are',groups)
     groups=[[element] for element in groups]
+    #print('groups converted as follows', groups)
+    #plot_clusters(groups)
+    #plt.show()
     for n in range(max_iterations):
         for i in range(data.shape[0]):
             group_no = classify_a_point(data[i,:], groups,k)
-            groups[group_no].append(data[i,:])
+            #print('point chosen: ', data[i,:], 'classified in group no. :', group_no)
+            print('data point is: ', data[i,:])
+            print('initialized point is: ',groups[group_no][0] )
+            if groups[group_no][0].all()!=data[i,:].all():  #to prevent the initialized points from gettind re-added
+                groups[group_no].append(data[i,:])
+            #plot_clusters(groups)
+            #plt.show()
     return groups
 
 def plot_clusters(groups):
-    colors = cm.rainbow(np.linspace(0, 1, len(groups)))
-    for group in range(len(groups)):
-        plt.scatter(*zip(*groups[group]), color=colors)
-    plt.show()
-
-plot_clusters(cluster(data, 4, 5,5))
+    #colors = cm.rainbow(np.linspace(0, 1, len(groups)))
+    plt.scatter(*zip(*groups[0]),[6], 'r')
+    plt.scatter(*zip(*groups[1]),[6], 'b')
+    plt.scatter(*zip(*groups[2]),[6], 'g')
+    plt.scatter(*zip(*groups[3]),[6], 'c')
+    plt.show()   
+
+
+
+l=initialize(data,4)
+colors = cm.rainbow(np.linspace(0, 1, len(l)))
+#cluster(data,4,5,1)
+plt.scatter(*zip(*data),[6],'r')
+for i in range(len(l)):
+    plt.scatter(l[i][0],l[i][1],[6],color=colors)
+plt.show()
+