【机器学习实战】10.K均值聚类算法

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @Time    : 18/4/15 下午4:32
# @Author  : cicada@hole
# @File    : kMeans.py
# @Desc    :  第10章 利用k均值聚类算法对未标注数据分组
# @Link    :

from numpy import *

'''
=========从文本导入数据
1.文本每行由tab分割
'''


def loadDataSet(fileName):
    dataMat = []
    fr = open(fileName)
    for line in fr.readlines():
        curLine = line.strip().split('\t')
        # print(curLine)
        fltLine = float(curLine[0]), float(curLine[1])
        print(fltLine)
        dataMat.append(fltLine)
    return dataMat


'''
======计算向量的欧式距离
(x1-x2)^2再开方
'''


def distEclud(vecA, vecB):
    return sqrt(sum(power(vecA - vecB, 2)))


'''
=======为给定数据集构建包含K个随机质心的集合

'''


def randCent(dataSet, k):
    numSamples, dim = matrix(dataSet).shape  # dim列数
    # n = dataSet.shape[1]
    n = dim
    centroids = matrix(zeros((k, n))) #k行n列
    for j in range(n):
        minJ = min(dataSet[:, j])
        rangeJ = float(max(dataSet[:, j]) - minJ)
        centroids[:, j] = minJ + rangeJ * random.rand(k, 1) #随机生成k行n列中心点
    return centroids


def kMeans(dataSet, k, distMeas=distEclud, createCent=randCent):
    m = shape(dataSet)[0]  # 样本数
    print("----样本数", m)
    clusterAssment = matrix(zeros((m, 2))) # mx2矩阵 簇分配矩阵 两列,一列簇索引值,一列存储误差

    centroids = createCent(dataSet, k) #随机选取中心
    clusterChanged = True  # 是否需要继续迭代
    while clusterChanged:
        clusterChanged = False
        for i in range(m): #遍历样本
            minDist = inf
            minIndex = -1
            for j in range(k):  # 寻找最近的质心
                distJI = distMeas(centroids[j, :], dataSet[i, :]) #计算样本和质心的距离
                if distJI < minDist:
                    minDist = distJI
                    minIndex = j
            if clusterAssment[i, 0] != minIndex:
                clusterChanged = True
            clusterAssment[i, :] = minIndex, minDist ** 2
        print(centroids)

        # 更新质心的位置
        for cent in range(k):
            ptsInClust = dataSet[nonzero(clusterAssment[:, 0].A == cent)[0]]
            centroids[cent, :] = mean(ptsInClust, axis=0)  # 计算均值
    return centroids, clusterAssment


'''
展示数据点
'''
import matplotlib.pyplot as plt
def showData(dataSet):
    x = []
    y = []
    plt.figure(figsize=(9, 6))
    for i in dataSet:
        x.append([float(i[0])])
        y.append([float(i[1])])
    plt.scatter(x, y, c='b', s = 25, alpha = 0.4, marker ='o')
    # c:散点的颜色
    # s：散点的大小
    # alpha:是透明程度
    plt.show()



def showCluster(dataSet, k, centroids, clusterAssment):
    numSamples, dim = dataSet.shape
    if dim != 2:
        print("Sorry! I can not draw because the dimension of your data is not 2!")
        return 1
    # v 倒三角 s正方形 o圆 +加号
    mark = ['vr', 'sb', 'og', '+k', '^r', '+r', 'sr', 'dr', '<r', 'pr']
    if k > len(mark):
        return 1

    # draw all samples
    for i in range(numSamples):
        markIndex = int(clusterAssment[i, 0])
        plt.plot(dataSet[i, 0], dataSet[i, 1], mark[markIndex])

    # D钻石
    mark = ['Dr', 'Db', 'Dg', 'Dk', '^b', '+b', 'sb', 'db', '<b', 'pb']
    # draw the centroids
    for i in range(k):
        plt.plot(centroids[i, 0], centroids[i, 1], mark[i], markersize=12)
    plt.show()

if __name__ == '__main__':
    print("======第一步: 加载数据====")
    dataMat = mat(loadDataSet('testSet.txt'))
    # showData(dataMat)
    k = 4
    myCentroids, clustAssing = kMeans(dataMat, k)
    print(myCentroids, clustAssing)

    showCluster(dataMat,k,myCentroids,clustAssing)