一.背景简要说明

二.模块代码及注释

from numpy import *
import operator#样本数据集创建函数
def creatDataSet():dataSet = array([[1.0,1.1],[1.0,1.0],[0,0],[0,0.1]])labels = ['A','A','B','B']return dataSet,labels#k-NN简单实现函数
def classify0(inX,dataSet,labels,k):#求出样本集的行数，也就是labels标签的数目dataSetSize = dataSet.shape[0]#构造输入值和样本集的差值矩阵diffMat = tile(inX,(dataSetSize,1)) - dataSet#计算欧式距离sqDiffMat = diffMat**2sqDistances = sqDiffMat.sum(axis=1)distances = sqDistances**0.5#求距离从小到大排序的序号sortedDistIndicies = distances.argsort()#对距离最小的k个点统计对应的样本标签classCount = {}for i in range(k):#取第i+1近邻的样本对应的类别标签voteIlabel = labels[sortedDistIndicies[i]]#以标签为key，标签出现的次数为value将统计到的标签及出现次数写进字典classCount[voteIlabel] = classCount.get(voteIlabel,0) + 1#对字典按value从大到小排序sortedClassCount = sorted(classCount.items(),key=operator.itemgetter(1),reverse=True)#返回排序后字典中最大value对应的keyreturn sortedClassCount[0][0]#从文本文件中解析数据
def file2matrix(filename):fr = open(filename)#按行读取整个文件的内容arrayOfLines = fr.readlines()#求文件行数，即样本总数numberOfLines = len(arrayOfLines)#初始化返回数组及列表returnMat = zeros((numberOfLines,3))classLabelVector = []index = 0for line in arrayOfLines:#截取掉每一行的回车符line = line.strip()#将一行数据根据制表符分割成包含多个元素（特征）的列表listFromLine = line.split('\t')#分割后的特征数据列表存入待返回的numpy数组returnMat[index,:] = listFromLine[0:3]#将文件每一行最后一列的数据存入待返回的分类标签列表classLabelVector.append(int(listFromLine[-1]))#控制待返回的numpy数组进入下一行index+=1return returnMat,classLabelVector#特征变量归一化
def autoNorm(dataSet):#取出每一列的最小值，即每一个特征的最小值minVals = dataSet.min(0)#取出每一列的最大值，即每一个特征的最大值maxVals = dataSet.max(0)#每一个特征变量的变化范围ranges = maxVals - minVals#初始化待返回的归一化特征数据集normDataSet = zeros(shape(dataSet))#特征数据集行数，即样本个数m = dataSet.shape[0]#利用tile()函数构造与原特征数据集同大小的矩阵，并进行归一化计算normDataSet = dataSet - tile(minVals,(m,1))normDataSet = normDataSet/tile(ranges,(m,1))return normDataSet,ranges,minVals#分类器测试
def datingClassTest():#给定用于测试分类器的样本比例hoRatio = 0.10#调用文本数据解析函数datingDataMat,datingLabels = file2matrix('F:/programming tools/datingTestSet2.txt')#调用特征变量归一化函数normMat,ranges,minVals = autoNorm(datingDataMat)#归一化的特征数据集行数，即样本个数m = normMat.shape[0]#用于测试分类器的测试样本个数numTestVecs = int(m*hoRatio)#初始化分类器犯错样本个数errorCount = 0.0for i in range(numTestVecs):#以测试样本作为输入，测试样本之后的样本作为训练样本，对测试样本进行分类classifierResult = classify0(normMat[i,:],normMat[numTestVecs:m,:],datingLabels[numTestVecs:m],3)#对比分类器对测试样本预测的类别和其真实类别print("the classifier came back with: %d,the real answer is: %d" % (classifierResult,datingLabels[i]))#统计分类出错的测试样本数if (classifierResult != datingLabels[i]):errorCount+=1.0#输出分类器错误率        print("the total error rate is: %f" % (errorCount/float(numTestVecs)))#约会网站对新输入分类
def classifyPerson():#定义一个存储了三个字符串的列表，分别对应不喜欢，一般喜欢，很喜欢resultList = ['not at all','in small dose','in large dose']#用户输入三个特征变量，并将输入的字符串类型转化为浮点型ffMiles = float(input("frequent flier miles earned per year:"))percentats = float(input("percentage of time spent playing video games:"))iceCream = float(input("liters of ice cream consumed per year:"))#调用文本数据解析函数datingDataMat,datingLabels = file2matrix('F:\programming tools\datingTestSet2.txt')#调用特征变量归一化函数normMat,ranges,minVals = autoNorm(datingDataMat)#将输入的特征变量构造成特征数组（矩阵）形式inArr = array([ffMiles,percentats,iceCream])#调用kNN简单实现函数classifierResult = classify0((inArr-minVals)/ranges,normMat,datingLabels,3)#将分类结果由数字转换为字符串print("You will probably like this person",resultList[classifierResult - 1])

三.详细解读

1.文本文件解析函数file2matrix