KNN-机器学习实战-Peter Harrington

#Machine Learning in Action: 机器学习实战, [美]Perer Harrington
from numpy import  *
import operator
import matplotlib
import matplotlib.pyplot as plt
from os import listdir
import time

#导入数据
def creatDataSet():
    group = array([[1.0,1.1],[1.0,1.0],[0,0],[0,0.1]])
    labels = ['A','A','B','B']
    return group, labels

#KNN分类算法
def classify0(inX, dataSet, labels, k):
    dataSetSize = dataSet.shape[0]
    diffMat = tile(inX, (dataSetSize,1)) - dataSet      #tile复制几份 inX 的值
    # print("diffMat:\n",diffMat)

    sqDiffMat = diffMat ** 2         #距离计算
    sqDistances = sqDiffMat.sum(axis=1)
    distances = sqDistances ** 0.5
    # print(distances)

    sortedDistIndicies = distances.argsort()       #返回下标,按距离最小到最大顺序排列, [2 3 1 0] ;;将元素从小到大排列，提取其对应的index(索引)
    # print(sortedDistIndicies)
    classCount = {}
    for i in range(k):       #选择距离最小的k个点,,疑问？？？
        voteIlabel = labels[sortedDistIndicies[i]]    #返回标签值--B B A
        classCount[voteIlabel] = classCount.get(voteIlabel, 0) + 1       #dict.get(key, default=None),返回指定键的值，如果值不在字典中返回默认值None。
        # print("voteIlabel:\n",classCount[voteIlabel])
    sortedClassCount = sorted(classCount.items(), key=operator.itemgetter(1),reverse =True)   #排序,reversed取值为True时候就是倒序排，默认为False正序从小到大排
                                                                                                # key：用列表元素的某个属性或函数进行作为关键字
    # print("sortedClassCount:\n",sortedClassCount)     #[('A', 1), ('B', 2)], 两种属性--A、B
    return sortedClassCount[0][0]

#将文本转换为Numpy能识别的格式
def file2matrix(filename):
    fr = open(filename)
    arrayOfLines = fr.readlines()
    numberOfLines = len(arrayOfLines)       #文件行数 ,"datingTestSet2.txt"有1000行，4列
    returnMat = zeros((numberOfLines, 3))   #创建返回的Numpy矩阵

    classLabelVector = []
    index = 0
    for line in arrayOfLines:
        line = line.strip()                   #截取所有的回车字符
        listFromLine = line.split('\t')       #使用tab字符将数据分割成元素列表
        returnMat[index,:] = listFromLine[0:3]   #三个数据
        classLabelVector.append(int(listFromLine[-1]))    #最后一列，标签类别
        index += 1
    return returnMat, classLabelVector
     # print(listFromLine[0:4])

#归一化特征值
def autoNorm(dataSet):
    minVals = dataSet.min(0)    #参数0表示从列中选取最小值
    maxVals = dataSet.max(0)
    ranges  = maxVals - minVals
    normDataSet = zeros(shape(dataSet))
    m = dataSet.shape[0]

    normDataSet = dataSet - tile(minVals, (m,1))
    normDataSet = normDataSet/tile(ranges, (m,1))
    return normDataSet, ranges, maxVals, minVals

#测试KNN模型
def datingClassTest():
    hoRatio = 0.10
    datingDataMat, datingLabels = file2matrix('datingTestSet2.txt')
    normMat, ranges, maxVals, minVals  = autoNorm(datingDataMat)
    m = normMat.shape[0]
    numTestVecs = int(m*hoRatio)     #1000*hoRatio

    errorCount  =0.0
    for i in range(numTestVecs):
        classfierResult = classify0(normMat[i,:], normMat[numTestVecs:m ,:],   #判断前面100行属于下面的900行的哪一类
                                    datingLabels[numTestVecs:m], 3)
        # print("classfier: %d, real: %d" % (classfierResult, datingLabels[i]))
        if (classfierResult != datingLabels[i]):
            errorCount += 1.0
    print("error_rate: %f" % (errorCount/float(numTestVecs)))
    print("acc_rate:   %f" % (1-(errorCount/float(numTestVecs))))
    # print(normMat)

#约会网站预测函数
def classifyPerson():
    resultList = ['not','small','large']
    percentTats = float(input("percentage of time?"))
    ffMiles = float(input("frequent flier miles earned?"))
    iceCream = float(input("liters of ice consumed?"))

    datingDataMat, datingLabels = file2matrix('datingTestSet2.txt')
    normMat, ranges, maxVals, minVals  = autoNorm(datingDataMat)
    inArr = array([ffMiles, percentTats, iceCream])

    classifierResult = classify0((inArr-minVals)/ranges, normMat, datingLabels, 3)
    print("you like person:", resultList[classifierResult - 1])       #为啥减去1 ？？
    print(classifierResult)

#图像转换为测试向量
def img2vector(filename):
    returnVector = zeros((1,1024))
    fr = open(filename)
    for i in range(32):
        lineStr = fr.readline()
        for j in range(32):
            returnVector[0,32*i+j] = int(lineStr[j])
            # print("returnVector:",returnVector[0,32*i:32*i+j])
    return returnVector

#手写数字识别系统测试
def handwritingClassTest():
    hwLabels = []
    trainingFileList = listdir('trainingDigits')
    m = len(trainingFileList)     #1934
    trainingMat = zeros((m, 1024))
    for i in range(m):
        fileNameStr = trainingFileList[i]
        # fileStr = fileNameStr.split('.')           # ['9_99', 'txt'] 等
        fileStr = fileNameStr.split('.')[0]         #split--指定分隔符对字符串进行切片，参数num仅分隔num个子字符串,指定字符不再出现。
        classNumStr = int(fileStr.split('_')[0])    # 输出 0, 1 ，2 ，3 ，...9
        hwLabels.append(classNumStr)
        trainingMat[i, :] = img2vector('trainingDigits/%s' % fileNameStr)

    testFileList = listdir('testDigits')
    errorCount = 0.0
    mTest = len(testFileList)    #946
    for i in range(mTest):
        fileNameStr = testFileList[i]
        fileStr = fileNameStr.split('.')[0]
        classNumStr = int(fileStr.split('_')[0])
        vectorUnderTest = img2vector('testDigits/%s' % (fileNameStr))
        classifierResult = classify0(vectorUnderTest, trainingMat, hwLabels, 3)
        print("classifier: %d, real_answer: %d" % (classifierResult, classNumStr))
        if (classifierResult != classNumStr):
            errorCount += 1.0

    print("error_num:",errorCount)
    print("error_rate:", (errorCount/float(mTest)))


if __name__ == '__main__':
    #创建数据
    group, labels = creatDataSet()
    print("group:\n", group)
    print("labels:\n",labels)
    #KNN算法
    sortedClassCount= classify0([0,0], group, labels, 3)
    # print("sortedClassCount[0][0]:\n",sortedClassCount)
    #文本转换为Numpy格式
    datingDataMat, datingLabels = file2matrix("datingTestSet2.txt")
    print("datingDataMat:\n",datingDataMat)
    # print("datingLabels:\n", datingLabels)
    # fig = plt.figure()          #Creates a new figure.
    # ax  = fig.add_subplot(111)
    # # ax.scatter(datingDataMat[:, 1], datingDataMat[:, 2])   #散点图，一种颜色
    # ax.scatter(datingDataMat[:, 1], datingDataMat[:, 2], 15.0*array(datingLabels), 15.0*array(datingLabels))   #散点图, 三种颜色
    # plt.show()

    #归一化特征值
    normMat, ranges, maxVals, minVals = autoNorm(datingDataMat)
    # print(normMat, ranges, maxVals, minVals)

    #测试代码
    datingClassTest()

    #约会网站预测函数
    # classifyPerson()

    #图像转换为测试向量
    testVector = img2vector('testDigits/1_13.txt')
    # print("testVector:\n", testVector[0,32*5:32*5+32])

    #手写数字识别系统测试
    star = time.time()
    handwritingClassTest()
    end = time.time()
    print("测试时间：%f s" % (end-star))   #测试时间：45.612609s