Python-常用机器学习算法-K-means

'''

创建k个点作为初始的质心点（随机选择）

当任意一个点的簇分配结果发生改变时

       对数据集中的每一个数据点

              对每一个质心

                     计算质心与数据点的距离

                    将数据点分配到距离最近的簇

            对每一个簇，计算簇中所有点的均值，并将均值作为质心

'''

def distEclud(vecA, vecB): #两个向量之间的距离

    return sqrt(sum(power(vecA - vecB, 2))) #la.norm(vecA-vecB) 欧几里得距离

def distSLC(vecA, vecB):#Spherical Law of Cosines 余弦距离(球面距离)

    a = sin(vecA[0,1]*pi/180) * sin(vecB[0,1]*pi/180)

    b = cos(vecA[0,1]*pi/180) * cos(vecB[0,1]*pi/180) * \

                      cos(pi * (vecB[0,0]-vecA[0,0]) /180)

    return arccos(a + b)*6371.0 #pi is imported with numpy

def randCent(dataSet, k): #创建k各随机点作为初始中心点

    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

def kMeans(dataSet, k, distMeas=distEclud, createCent=randCent):

    m = shape(dataSet)[0]

    clusterAssment = mat(zeros((m,2)))#create mat to assign data points 

                               #to a centroid, also holds SE of each point

    centroids = createCent(dataSet, k)

    clusterChanged = True

    while clusterChanged:

        clusterChanged = False

        for i in range(m):   #for each data point assign it to the closest centroid

            minDist = inf

            minIndex = -1

            for j in range(k):

                distJI = distMeas(centroids[j,:],dataSet[i,:])

                if distJI < minDist:

                    minDist = distJI; minIndex = j

            if clusterAssment[i,0] != minIndex: clusterChanged = True

            clusterAssment[i,:] = minIndex,minDist**2

        # print centroids

        for cent in range(k):#recalculate centroids

            ptsInClust = dataSet[nonzero(clusterAssment[:,0].A==cent)[0]]#get all the point in this cluster

            centroids[cent,:] = mean(ptsInClust, axis=0) #assign centroid to mean

    return centroids, clusterAssment