第10章 利用K-均值聚类算法对未标注数据分组

程序清单10-1 K-均值聚类支持函数

import matplotlib
from time import sleep
import urllib
import matplotlib.pyplot as plt
import json
from numpy import *


def loadDataSet(fileName):
    """
    输入：将文本文件导入到一个列表中
    返回值：一个包含许多其他列表的列表
    """
    dataMat = []  # assume last column is target value
    fr = open(fileName)
    for line in fr.readlines():
        curLine = line.strip().split('\t')
        fltLine = list(map(float, curLine))  # map all elements to float()
        dataMat.append(fltLine)
    return dataMat


def distEclud(vecA, vecB):
    """
    计算两个向量的欧氏距离
    """
    return sqrt(sum(power(vecA - vecB, 2)))


def randCent(dataSet, k):
    """
    为给定数据集构建一个包含k个随机质心的集合,质心大小要在min和max之间。
    """
    n = shape(dataSet)[1]
    centroids = mat(zeros((k, n)))  # create centroid mat
    for j in range(n):  # create random cluster centers, within bounds of each dimension
        minJ = min(dataSet[:, j])
        rangeJ = float(max(dataSet[:, j]) - minJ)
        centroids[:, j] = mat(minJ + rangeJ * random.rand(k, 1))
    return centroids

# 测试(逐条)
# datMat = mat(kMeans.loadDataSet('testSet.txt'))
# datMat
# min(datMat[:, 0])
# min(datMat[:, 1])
# max(datMat[:, 0])
# max(datMat[:, 1])
# kMeans.randCent(datMat, 2)  # 两个质心
# kMeans.distEclud(datMat[0], datMat[1])  # 欧氏距离

代码清单10-3 二分k-均值聚类算法

def biKmeans(dataSet, k, distMeas=distEclud):
    """
    二分k-均值算法
    返回值：质心列表、簇分配结果
    """
    m = shape(dataSet)[0]
    clusterAssment = mat(zeros((m, 2)))  # 存储数据集中每个点的簇分配结果及平方误差
    centroid0 = mean(dataSet, axis=0).tolist()[0]  # 计算整个数据集的质心
    centList = [centroid0]  # 使用一个列表来保留所有的质心
    for j in range(m):  # 计算每个点到质心的误差值
        clusterAssment[j, 1] = distMeas(mat(centroid0), dataSet[j, :])**2
    while (len(centList) < k):
        lowestSSE = inf
        for i in range(len(centList)):  # 遍历所有的簇
            # get the data points currently in cluster i
            ptsInCurrCluster = dataSet[nonzero(  # 次数据集
                clusterAssment[:, 0].A == i)[0], :]
            centroidMat, splitClustAss = kMeans(
                ptsInCurrCluster, 2, distMeas)  # k=2，生成两个质心和个每个簇的误差值
            # compare the SSE to the currrent minimum
            sseSplit = sum(splitClustAss[:, 1])
            sseNotSplit = sum(clusterAssment[nonzero(
                clusterAssment[:, 0].A != i)[0], 1])
            print("sseSplit, and notSplit: ", sseSplit, sseNotSplit)
            if (sseSplit + sseNotSplit) < lowestSSE:  # 这些误差与剩余数据集的误差之和作为本次划分的误差
                bestCentToSplit = i
                bestNewCents = centroidMat
                bestClustAss = splitClustAss.copy()
                lowestSSE = sseSplit + sseNotSplit
        # 当使用kMeans()函数并且指定簇数为2时，会得到两个编号分别为0和1的结果簇。需要将这些簇编号修改为划分簇
        # 及新加簇的编号，该过程可以通过两个数组过滤器来完成。
        bestClustAss[nonzero(bestClustAss[:, 0].A == 1)[0], 0] = len(
            centList)  # change 1 to 3,4, or whatever
        bestClustAss[nonzero(bestClustAss[:, 0].A == 0)
                     [0], 0] = bestCentToSplit
        print('the bestCentToSplit is: ', bestCentToSplit)
        print('the len of bestClustAss is: ', len(bestClustAss))
        # replace a centroid with two best centroids
        centList[bestCentToSplit] = bestNewCents[0, :].tolist()[
            0]  # 新的簇分配结果被更新
        centList.append(bestNewCents[1, :].tolist()[0])  # 新的质心会被添加到列表centList中
        clusterAssment[nonzero(clusterAssment[:, 0].A == bestCentToSplit)[
            0], :] = bestClustAss  # reassign new clusters, and SSE
    return mat(centList), clusterAssment


# 测试代码
# import kMeans
# datMat3 = mat(kMeans.loadDataSet('testSet2.txt'))
# centList,myNewAssments = kMeans.biKmeans(datMat3,3)
# centList