代码注释：机器学习实战第2章 k-近邻算法

写在开头的话：在学习《机器学习实战》的过程中发现书中很多代码并没有注释，这对新入门的同学是一个挑战，特此贴出我对代码做出的注释，仅供参考，欢迎指正。

1、导入数据：

#coding:gbk
from numpy import *
import operator

def createDataSet():
    group = array([[1.0, 1.1], [1.0, 1.0], [0, 0],[0, 0.1]])#4*2维矩阵
    labels  = ['A', 'A', 'B', 'B']
    return group, labels

注释：有中文注释必须加

#coding:gbk

2、k-近邻算法

def classify0(inX, dataSet, labels, k):
    dataSetSize = dataSet.shape[0]#dataSet一维长度
    diffMat = tile(inX, (dataSetSize, 1)) - dataSet#将inx扩展成dataSet一样矩阵后相减
    sqDiffMat = diffMat**2#平方
    sqDistances = sqDiffMat.sum(axis = 1)#按行求和
    distances = sqDistances**0.5#开根号
    sortedDistIndicies = distances.argsort()#返回distances从小到大的索引值
    classCount = {}#建立字典，用于指示labels数多少
    for i in range(k):
        voteIlabel = labels[sortedDistIndicies[i]]#返回第i个label值
        classCount[voteIlabel] = classCount.get(voteIlabel,0) + 1#对每个label值出现的频率计数，0代表第一次计数，字典对应出现的字数为0
    #对classCount按照值的大小从大到小进行排序，返回list
    sortedClassCount = sorted(classCount.iteritems(),#iteritems()表示将classCount以一个迭代器对象返回
                              key = operator.itemgetter(1), reverse = True)#operator.itemgetter(1)表示第2维数据即值，reverse = True表示从大大小排列
    return sortedClassCount[0][0]

3、文本转换为Numpy矩阵

def file2matrix(filename):
    fr = open(filename)
    arrayOLines = fr.readlines()
    numberOfLines = len(arrayOLines)#文件行数
    returnMat = zeros((numberOfLines,3))#创建空矩阵，表示训练样本矩阵
    classLabelVector = []#创建空列表，表示类标签向量
    index = 0#表示第index行
    for line in arrayOLines:
        line = line.strip()#删除空白符（包括'\n', '\r',  '\t',  ' ')
        listFromLine = line.split('\t')#以'\t'分割字符串
        returnMat[index,:] = listFromLine[0:3]#训练样本矩阵赋值
        classLabelVector.append(int(listFromLine[-1]))#类标签向量赋值
        index += 1
    return returnMat, classLabelVector

4、归一化特征值

def autoNorm(dataSet):
    minVals = dataSet.min(0)#得每列最小值，返回1*m数组
    maxVals = dataSet.max(0)#得每列最大值，返回1*m数组
    ranges = maxVals - minVals
    normDataSet = zeros(shape(dataSet))#和dataSet一样维度的空矩阵
    m = dataSet.shape[0]#得dataSet行数
    normDataSet = dataSet - tile(minVals, (m, 1))
    normDataSet = normDataSet/tile(ranges, (m, 1))
    return normDataSet, ranges, minVals

5、约会网站测试代码

def datingClassTest():
    hoRatio = 0.10#测试数据占10%
    datingDataMat, datingLabels = file2matrix('datingTestSet2.txt')#导入数据
    normMat, ranges, minVals = autoNorm(datingDataMat)#归一化数值
    m = normMat.shape[0]#数据总量
    numTestVecs = int(m*hoRatio)#测试数据量
    errorCount = 0.0#错误分类计数变量
    for i in range(numTestVecs):#对测试集中数据
        classifierResult = classify0(normMat[i, :], normMat[numTestVecs:m, :],datingLabels[numTestVecs:m], 3)
        if (classifierResult != datingLabels[i]):
            errorCount += 1.0
            print "the classfier came back with: %d, the real anwser is: %d" % (classifierResult, datingLabels[i])
    print "the total error rate is: %f%%" % (100 * errorCount / float (numTestVecs))

6、约会网站预测函数

def classifyPerson():
    resultList = ['not at all', 'in small doses', 'in large doses']
    percentTats = float(raw_input("percentage of time spent playing video games?"))#输入...时间百分比
    ffMiles = float(raw_input("frequent flier miles earned per year?"))#输入飞行公里数
    iceCream = float(raw_input("liters of ice cream consumed per year?"))#输入冰激凌量
    datingDataMat, datingLabels = file2matrix('datingTestSet2.txt')#导入数据
    normMat, ranges, minVals = autoNorm(datingDataMat)#归一化数值
    inArr = array([ffMiles, percentTats, iceCream])#需要预测数据
    classifierResult = classify0((inArr - minVals) / ranges, normMat, datingLabels, 3)
    print "You will probably like this person: ", resultList[classifierResult - 1]

7、图像转换为Numpy矩阵

def img2vector(filename):
    returnVect = zeros((1, 1024))
    fr = open(filename)
    for i in range(32):
        lineStr = fr.readline()
        for j in range(32):
            returnVect[0, 32 * i + j] = int (lineStr[j])
    return returnVect

8、手写数字识别系统测试代码

def handwritingClassTest():
    hwLabels = []#表示trainingMat代表的值，即类标签向量
    trainingFileList = listdir('trainingDigits')#得trainingDigits文件夹里的文件名
    m = len(trainingFileList)#trainingDigits文件夹里的文件数
    trainingMat = zeros((m, 1024))#训练样本矩阵
    for i in range(m):
        fileNameStr = trainingFileList[i]
        fileStr = fileNameStr.split('.')[0]#文件名里去除.txt
        classNumStr = int(fileStr.split('_')[0])#文件名里去除_i
        hwLabels.append(classNumStr)#类标签向量赋值
        trainingMat[i, :] = img2vector('trainingDigits/%s' % fileNameStr)#训练样本矩阵赋值
    testFileList = listdir('testDigits')#得testDigits文件夹里的文件名
    errorCount = 0.0#错误分类计数变量
    mTest = len(testFileList)#testDigits文件夹里的文件数，即测试数据量
    for i in range(mTest):
        fileNameStr = testFileList[i]
        fileStr = fileNameStr.split('.')[0]#文件名里去除.txt
        classNumStr = int(fileStr.split('_')[0])#文件名里去除_i
        vectorUnderTest = img2vector('testDigits/%s' % fileNameStr)#测试样本矩阵赋值
        classifierResult = classify0(vectorUnderTest, trainingMat, hwLabels, 3)
        if (classifierResult != classNumStr):
            errorCount += 1.0
            print "%d: the classfier came back with: %d, the real anwser is: %d" % (errorCount, classifierResult, classNumStr)
    print "\nthe total error rate is: %f %%" % (100* errorCount / float(mTest))