KNN海伦约会

记录一下knn经典算法，100行代码50行注释的那种，直接copy就能执行成功

from numpy import *
import operator


# 数据读取
def file2matrix(filename):
    fr = open(filename)
    numberOfLines = len(fr.readlines())
    returnMat = zeros((numberOfLines, 3))  # 生成对应行数的零矩阵；returnMats是返回的数据集列表
    classLabelVector = []  # return的标签列表
    fr = open(filename)
    index = 0
    for l in fr.readlines():
        # 以行为单位删除前后空格
        line = l.strip()
        # 以\t切割字符串,前三列存入数据集列表中
        listFormLine = line.split('\t')
        # listFormLine中[0，3)个元素放入returnMat第index行中
        returnMat[index, :] = listFormLine[0:3]
        # 最后一列标签列存到标签列表中
        classLabelVector.append(int(listFormLine[-1]))
        index += 1
    return returnMat, classLabelVector


# 数据集归一化
def autoNorm(dataSet):
    '''
    dataSet：数据集
    归一化公式：Y=(X-Xmin)/(Xmax-Xmin)
    '''
    # 求出最大值，最小值，范围
    minVals = dataSet.min(0)
    maxVals = dataSet.max(0)
    ranges = maxVals - minVals
    normDataSet = zeros(shape(dataSet))  # shape(array)看array是几行几列
    m = dataSet.shape[0]  # shape[0]查看行，shape[1]查看列
    # tile(minVal,(m,1)将minVal重复m行1列
    normDataSet = dataSet - tile(minVals, (m, 1))  # 减去最小值
    normDataSet = normDataSet / tile(ranges, (m, 1))  # 除以范围
    return normDataSet, ranges, minVals


# KNN算法,X是输入的归一化之后的数据，dataSet是归一化后的数据集(3列)
def classify0(intX, dataSet, labels, k):
    dataSetSize = dataSet.shape[0]
    # 计算欧氏距离：所有数据都与输入的算(把输入的数据重复dataSetSize行)
    diffMat = tile(intX, (dataSetSize, 1)) - dataSet
    sqDiffMat = diffMat ** 2
    sqDistance = sqDiffMat.sum(axis=1)  # 将矩阵每一行中的向量相加：[[0,1,2] [2,1,3]]-》[3,6]
    distance = sqDistance ** 0.5
    # 排序，argsort()用法，表示对数据进行从小到大进行排序，返回数据的索引值。
    sortedDistIndicies = distance.argsort()
    # 选K个，判断所属类别
    classCount = {}  # 字典，统计每个标签的出现次数
    for i in range(k):
        voteIlabel = labels[sortedDistIndicies[i]]
        # classCount.get(voteIlabel, 0)在字典里找voteIlabel，没有就默认0，有几就是几
        classCount[voteIlabel] = classCount.get(voteIlabel, 0) + 1  # 出现次数加1
    # sorted(可迭代对象，自定义排序逻辑，True为降序False为升序（默认）)
    # dict.items()以列表返回可遍历的(键, 值) 元组数组
    # operator.itemgetter(0)按字典的key排序；operator.itemgetter(1)按value排序
    sortedClassCount = sorted(classCount.items(), key=operator.itemgetter(1), reverse=True)
    return sortedClassCount[0][0]


# KNN测试
def datingClassTest():
    hoRatio = 0.1  # 测试集数据与训练集数据比例
    # 调用数据读取函数获取数据集和标签列
    datingDataMat, datingLabels = file2matrix('datingTestSet2.txt')
    # 调用数据归一化函数，归一化数据
    normMat, ranges, minVals = autoNorm(datingDataMat)
    m = normMat.shape[0]  # m为数据的行数即矩阵的第一维
    numTestVecs = int(m * hoRatio)  # 测试样本的数量
    print("numTestVecs=", numTestVecs)
    errorCount = 0.0
    for i in range(numTestVecs):
        classfierResult = classify0(normMat[i, :], normMat[numTestVecs:m, :], datingLabels[numTestVecs:m], 3)
        print("the classifier came back with:%d,the real answer is:%s"%(classfierResult, datingLabels[i]))
        if classfierResult != datingLabels[i]:
            errorCount += 1
        print("total error rate is:", errorCount / float(numTestVecs))
        print("errorCount:", errorCount)


# 约会网站预测函数
def classifyPerson():
    resultList = ['not at all', 'in small doses', 'in large doses']
    percentTats = float(input("percentage of time spent playing video games?"))
    ffMiles = float(input("flier miles earned per year?"))
    iceCream = float(input("liters of ice cream consumed per year?"))
    datingDataMat, datingLabels = file2matrix('datingTestSet2.txt')
    normMat, ranges, minVals = autoNorm(datingDataMat)
    inArr = array([ffMiles, percentTats, iceCream])
    # 将输入的数据归一化后传入knn算法函数
    classifierResult = classify0((inArr - minVals) / ranges, normMat, datingLabels, 3)
    print("you will probably like this person:", resultList[classifierResult - 1])


if __name__ == '__main__':
    #classifyPerson()
    datingClassTest()