文章给出层次聚类算法的python实现方法，并用《数据挖掘导论》里面的具体数据进行运行，代码如下：

from numpy import *
from math import *
from operator import *
def dist(a,b):#a,b is mat
    c=(a-b)*(a-b).T
    return sqrt(sum(c))
    
def centroid(a): #a is mat
    return a.mean(0)

def resolveList(List):#分解一个List，如将[[1,2],[[1,2],[1,2]]]分解为[[1,2],[1,2],[1,2]]
    L=[]
    for i in range(len(List)):
        if type(List[i][0])!=list:
            L.append(List[i])
            continue
        else:
            List1=resolveList(List[i])
            for j in List1:
                L.append(j)
    return L


def cluster(data,num): #data is list   num is the number of clusters you want
    dataList=data
    clusterList=range(len(dataList))
    while len(dataList)>num:
        dataMat=map(mat,data)
        distDic={}
        for i in range(len(dataMat)-1):
            for j in range(len(dataMat))[i+1:]:
                distDic[(i,j)]=dist(dataMat[i].mean(0),dataMat[j].mean(0))
        m,n=sorted(distDic.items(),key=itemgetter(1))[0][0]
        tempList=[]
        tempList.append(dataList[m])
        tempList.append(dataList[n])
        L=resolveList(tempList)
        del dataList[n]
        del dataList[m]
        dataList.append(L)
        print dataList  #打出每一次结果

data=[[0.4005,0.5306],[0.2148,0.3854],[0.3457,0.3156],[0.2652,0.1875],[0.0789,0.4139],[0.4548,0.3022]]
cluster(data,2)

结果如下：

[[0.4005, 0.5306], [0.2148, 0.3854], [0.2652, 0.1875], [0.0789, 0.4139], [[0.3457, 0.3156], [0.4548, 0.3022]]]
[[0.4005, 0.5306], [0.2652, 0.1875], [[0.3457, 0.3156], [0.4548, 0.3022]], [[0.2148, 0.3854], [0.0789, 0.4139]]]
[[0.4005, 0.5306], [[0.2148, 0.3854], [0.0789, 0.4139]], [[0.2652, 0.1875], [0.3457, 0.3156], [0.4548, 0.3022]]]
[[0.4005, 0.5306], [[0.2148, 0.3854], [0.0789, 0.4139], [0.2652, 0.1875], [0.3457, 0.3156], [0.4548, 0.3022]]]

程序可能编写的很菜，如有错误，希望指正~