Python实现K-means聚类算法

 1 from numpy import *  
 2 import time  
 3 import matplotlib.pyplot as plt  
 4   
 5   
 6 # 判读两点之间的距离  
 7 def euclDistance(x1, x2):  
 8     return sqrt(sum(power(x2 - x1, 2)))  
 9   
10 # 初始化质心与随机样本点 
11 def initCentroids(data, k):
12     len_line, dim = data.shape  #读取矩阵的长度,返回一个（len_line，dim）的元组
13     shuzu = zeros((k, dim))  #创建k行dim列的数组
14     for i in range(k):  
15         index = int(random.uniform(0, len_line)) #uniform()方法将随机生成下一个实数，它在[x,y]范围内。
16         shuzu[i, :] = data[index, :]  
17     return shuzu  
18   
19 # k-means cluster  
20 def kmeans(data, k):  
21     len_line = data.shape[0]  #读取数组第一行的长度
22     
23    
24     clusterAssment = mat(zeros((len_line, 2)))  
25     clusterChanged = True  
26   
27     ## step 1: init centroids  
28     centroids = initCentroids(data, k)  
29   
30     while clusterChanged:  
31         clusterChanged = False  
32         ## for each sample  
33         for i in xrange(len_line):  
34             minDist  = 100000.0  
35             minIndex = 0  
36             ## for each centroid  
37             ## step 2: find the centroid who is closest  
38             for j in range(k):  
39                 distance = euclDistance(centroids[j, :], data[i, :])  
40                 if distance < minDist:  
41                     minDist  = distance  
42                     minIndex = j  
43               
44             ## step 3: update its cluster  
45             if clusterAssment[i, 0] != minIndex:  
46                 clusterChanged = True  
47                 clusterAssment[i, :] = minIndex, minDist**2  
48   
49         ## step 4: update centroids  
50         for j in range(k):  
51             pointsInCluster = data[nonzero(clusterAssment[:, 0].A == j)[0]]  
52             centroids[j, :] = mean(pointsInCluster, axis = 0)  
53   
54     print ‘Congratulations, cluster complete!‘  
55     return centroids, clusterAssment  
56   
57 # show your cluster only available with 2-D data  
58 def showCluster(data, k, centroids, clusterAssment):  
59     len_line, dim = data.shape  
60     if dim != 2:  
61         print "Sorry! I can not draw because the dimension of your data is not 2!"  
62         return 1  
63   
64     mark = [‘or‘, ‘ob‘, ‘og‘, ‘ok‘, ‘^r‘, ‘+r‘, ‘sr‘, ‘dr‘, ‘<r‘, ‘pr‘]  
65     if k > len(mark):  
66         print "Sorry! Your k is too large! please contact Zouxy"  
67         return 1  
68   
69     # draw all samples  
70     for i in xrange(len_line):  
71         markIndex = int(clusterAssment[i, 0])  
72         plt.plot(data[i, 0], data[i, 1], mark[markIndex])  
73   
74     mark = [‘Dr‘, ‘Db‘, ‘Dg‘, ‘Dk‘, ‘^b‘, ‘+b‘, ‘sb‘, ‘db‘, ‘<b‘, ‘pb‘]  
75     # draw the centroids  
76     for i in range(k):  
77         plt.plot(centroids[i, 0], centroids[i, 1], mark[i], markersize = 12)  
78   
79     plt.show()

#################################################  
# kmeans: k-means cluster  
# Author : zouxy  
# Date   : 2013-12-25  
# HomePage : http://blog.csdn.net/zouxy09  
# Email  : zouxy09@qq.com  
#################################################  
  
from numpy import *  
import time  
import matplotlib.pyplot as plt  
  
## step 1: load data  
print "step 1: load data..."  
dataSet = []  
fileIn = open(‘E:/Python/Machine Learning in Action/testSet.txt‘)  
for line in fileIn.readlines():  
    lineArr = line.strip().split(‘\t‘)  
    dataSet.append([float(lineArr[0]), float(lineArr[1])])  
  
## step 2: clustering...  
print "step 2: clustering..."  
dataSet = mat(dataSet)  
k = 4  
centroids, clusterAssment = kmeans(dataSet, k)  
  
## step 3: show the result  
print "step 3: show the result..."  
showCluster(dataSet, k, centroids, clusterAssment)