python kmeans聚类简单介绍和实现代码

时间:2022-09-04 15:57:37

一、k均值聚类的简单介绍

假设样本分为c类,每个类均存在一个中心点,通过随机生成c个中心点进行迭代,计算每个样本点到类中心的距离(可以自定义、常用的是欧式距离)  

将该样本点归入到最短距离所在的类,重新计算聚类中心,进行下次的重新划分样本,最终类中心不改变时,聚类完成   

二、伪代码  

三、python代码实现  

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
#!/usr/bin/env python
# coding=utf-8
 
import numpy as np
import random
import matplotlib.pyplot as plt
 
#data:numpy.array dataset
#k the number of cluster
def k_means(data,k):
   
  #random generate cluster_center
  sample_num=data.shape[0]
  center_index=random.sample(range(sample_num),k)
  cluster_cen=data[center_index,:]
 
  is_change=1
  cat=np.zeros(sample_num)
   
 
  while is_change:
    is_change=0
 
    for i in range(sample_num):
      min_distance=100000
      min_index=0
 
      for j in range(k):
        sub_data=data[i,:]-cluster_cen[j,:]
        distance=np.inner(sub_data,sub_data)
        if distance<min_distance:
          min_distance=distance
          min_index=j+1
 
      if cat[i]!=min_index:
        is_change=1
        cat[i]=min_index
    for j in range(k):
      cluster_cen[j]=np.mean(data[cat==(j+1)],axis=0)
 
  return cat,cluster_cen
 
 
if __name__=='__main__':
 
  #generate data
  cov=[[1,0],[0,1]]
  mean1=[1,-1]
  x1=np.random.multivariate_normal(mean1,cov,200)
 
  mean2=[5.5,-4.5]
  x2=np.random.multivariate_normal(mean2,cov,200)
 
  mean3=[1,4]
  x3=np.random.multivariate_normal(mean3,cov,200)
 
  mean4=[6,4.5]
  x4=np.random.multivariate_normal(mean4,cov,200)
 
  mean5=[9,0.0]
  x5=np.random.multivariate_normal(mean5,cov,200)
   
  X=np.vstack((x1,x2,x3,x4,x5))
   
  #data distribution
  fig1=plt.figure(1)
  p1=plt.scatter(x1[:,0],x1[:,1],marker='o',color='r',label='x1')
  p2=plt.scatter(x2[:,0],x2[:,1],marker='+',color='m',label='x2')
  p3=plt.scatter(x3[:,0],x3[:,1],marker='x',color='b',label='x3')
  p4=plt.scatter(x4[:,0],x4[:,1],marker='*',color='g',label='x4')
  p5=plt.scatter(x5[:,0],x4[:,1],marker='+',color='y',label='x5')
  plt.title('original data')
  plt.legend(loc='upper right')
   
  cat,cluster_cen=k_means(X,5)   
 
  print 'the number of cluster 1:',sum(cat==1)
  print 'the number of cluster 2:',sum(cat==2)
  print 'the number of cluster 3:',sum(cat==3)
  print 'the number of cluster 4:',sum(cat==4)
  print 'the number of cluster 5:',sum(cat==5)
 
   
  fig2=plt.figure(2)
  for i,m,lo,label in zip(range(5),['o','+','x','*','+'],['r','m','b','g','y'],['x1','x2','x3','x4','x5']):
 
    p=plt.scatter(X[cat==(i+1),0],X[cat==(i+1),1],marker=m,color=lo,label=label)
  plt.legend(loc='upper right')
  plt.title('the clustering result')
  plt.show()

以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持服务器之家。

原文链接:http://blog.csdn.net/Jason____zhou/article/details/50283035