kNN算法实例（约会对象喜好预测和手写识别）

import numpy as np
import operator
import random
import os

def file2matrix(filePath):#从文本中提取特征矩阵和标签
    f = open(filePath,'r+').readlines()
    fileLength = len(f)
    dataSet = np.zeros((fileLength,3),np.float64)
    labelList = []
    for i in range(fileLength):
        row = f[i].split('\t')
        dataSet[i,:] = row[0:3]
        labelList.append(row[-1].strip('\n'))
    return dataSet,labelList

def autoNormal(data):#归一化处理
    dataShape = data.shape
    dataMin = data.min(0)
    dataMax = data.max(0)
    normalDataSet = np.zeros(dataShape,np.float64)
    diff = dataMax - dataMin
    normalDataSet = (data -np.tile(dataMin,(dataShape[0],1)))/np.tile(diff,(dataShape[0],1))
    return normalDataSet,diff,dataMin

def dataClassTest(dataSet,labelList):#测试算法准确率
    ratio = 0.1
    correntCount = 0
    testNumber = int(ratio*dataSet.shape[0])
    for i in range(testNumber):
        k = random.randint(0, dataSet.shape[0])
        label = classify0(dataSet[k],dataSet,labelList,20)
        if label == labelList[k]:
            correntCount += 1
    return correntCount*100/testNumber

def classifyPerson():#输入数据进行预测
    dataSet,labelSet = file2matrix('datingTestSet.txt')
    percentTats = float(input('Please input percentage of time spend playing video games?'))
    miles = float(input('Please input frequent flier miles earned per year?'))
    cream = float(input('Please input liters of ice cream consumed per year?'))
    dataSet,diff,dataMin = autoNormal(dataSet)
    intX = np.array([percentTats,miles,cream],np.float64)

    label = classify0((intX-dataMin)/diff,dataSet,labelSet,20)
    print("You likely {0} the man！".format(label))

    correntPercent = dataClassTest(dataSet,labelSet)
    print("The estimate corrent percent is {0}%！".format(correntPercent))

def classify0(intX,dataSet,labelSet,k):#kNN分类算法
    intX = np.tile(intX,(dataSet.shape[0],1))
    square = (intX - dataSet)**2
    sum = square.sum(axis=1)
    sqrt = sum**0.5
    sortedDistIndicies = sqrt.argsort()
    classCount={}
    for i in range(k):
        label = labelSet[sortedDistIndicies[i]]
        classCount[label] = classCount.get(label,0)+1
    sortedClassCount = sorted(classCount.items(),key=operator.itemgetter(1),reverse=True)

    return sortedClassCount[0][0]

def img2vector(filename):#将32*32图片转换成1*1024向量
    vector = np.zeros((1,1024))
    f = open(filename)
    for i in range(32):
        fr = f.readline()
        for j in range(32):
            vector[0,32*i+j] = int(fr[j])
    return vector

def handwritingClassTest():
    filenameList = os.listdir(r'machinelearninginaction\Ch02\digits\trainingDigits')
    m = len(filenameList)
    trainLabelList = []
    trainDataMatrix = np.zeros((m,1024))
    for i in range(m):
        trainLabelList.append(int(filenameList[i].strip('_')[0]))
        trainDataMatrix[i,:] = img2vector(r'machinelearninginaction\Ch02\digits\trainingDigits\{0}'.format(filenameList[i]))
    filenameList = os.listdir(r'machinelearninginaction\Ch02\digits\testDigits')
    m = len(filenameList)
    corrent = 0.0
    for i in range(m):
        testLabel = int(filenameList[i].strip('_')[0])
        testIn = img2vector(r'machinelearninginaction\Ch02\digits\testDigits\{0}'.format(filenameList[i]))
        testOut = classify0(testIn,trainDataMatrix,trainLabelList,3)
        if testOut == testLabel:
            corrent += 1
        else:
            print("Error:the classifier came back with:{0}, the real answer is:{1}。".format(testOut,testLabel))
    print("the corrent percent is:%.2f %%。"%(corrent*100/m))
if __name__ == '__main__':
    classifyPerson() #约会预测
    #handwritingClassTest() #手写识别

约会预测运行结果：

Please input percentage of time spend playing video games?100
Please input frequent flier miles earned per year?8
Please input liters of ice cream consumed per year?200
You likely didntLike the man！
The estimate corrent percent is 96.0%！

进程已结束，退出代码 0

手写识别运行结果：

Error:the classifier came back with:7, the real answer is:1。
Error:the classifier came back with:9, the real answer is:3。
Error:the classifier came back with:3, the real answer is:5。
Error:the classifier came back with:6, the real answer is:5。
Error:the classifier came back with:6, the real answer is:8。
Error:the classifier came back with:3, the real answer is:8。
Error:the classifier came back with:1, the real answer is:8。
Error:the classifier came back with:1, the real answer is:8。
Error:the classifier came back with:1, the real answer is:9。
Error:the classifier came back with:7, the real answer is:9。
the corrent percent is:98.94 %。

进程已结束，退出代码 0

测试数据：

说明：代码参考《机器学习实战》