机器学习实战——K-近邻算法（读书笔记）

K-近邻算法（Python）

def classify0(inx, dataSet, labels, k):
    """
    :param inx: 测试样本集合
    :param dataSet: 训练样本集合
    :param labels: 标签集合
    :param k: 选取k个相似度最高
    :return: 返回测试样本标签
    """
    #距离计算
    dataSetSize = dataSet.shape[0] #返回数组大小 group为4*2 shape=（4,2）
    diffMax = np.tile(inx, (dataSetSize,1)) -dataSet #tile 根据训练数据大小生成 矩阵减法
    sqDiffMax = diffMax**2 #平方
    sqDistances = sqDiffMax.sum(axis=1) #行坐标累加
    distances = sqDistances**0.5 #开根号

    #排序
    sortedDisIndicies = distances.argsort() #排序从小到大 返回索引

    #获取距离最小的K个点
    classCount = {}
    for i in range(k):
        voteIlabel = labels[sortedDisIndicies[i]] #根据下标获取标签
        classCount[voteIlabel] = classCount.get(voteIlabel,0)+1 #累计标签集合次数

    #集合K根据次数排序
    sortedClassCount = sorted(classCount.items(), key = operator.itemgetter(1), reverse = True)

    return sortedClassCount[0][0]

构造训练样本集合 标签集合

def createDataSet():
    group = np.array([[1.0,1.1],[1.0,1.0],[0,0],[0,0.1]])
    labels = ['A', 'A', 'B','B']
    return group, labels

测试调用

group,labels = knn.createDataSet()
res = knn.classify0([0,0],group,labels,3)

文件读取：1000行数据 前三列表示训练样本 最后一列为标签

def file2maxtrix(filename):
    fr = open(filename)
    arrayOLines = fr.readlines() 
    numberOfLines = len(arrayOLines) #获取文本行数
    returnMax = np.zeros((numberOfLines, 3))#生成1000*3 的 零矩阵
    classLabelVector = []
    index = 0
    for line in arrayOLines:
        line = line.strip()
        listFromLine = line.split('\t')
        returnMax[index,:] = listFromLine[0:3] #前三列是训练样本数据
        classLabelVector.append(int(listFromLine[-1])) #最后一列是标签集合
        index += 1
    return returnMax, classLabelVector

读取文件画图，在数据处理之前，通过图形化分析数据特征 数据通过https://www.manning.com/books/machine-learning-in-action sourcecode下载

filname = 'C:/javacode/machinelearninginaction/Ch02/datingTestSet2.txt'
datingDateMat, datingLabels = knn.file2maxtrix(filname)
print('datingDateMat: {}'.format(datingDateMat))

fig = plt.figure()
ax = fig.add_subplot(111)
ax.scatter(datingDateMat[:,0], datingDateMat[:,1], c=datingLabels)#c这个参数是表示颜色 根据标签分不同颜色展示 datingLabels数组 
plt.show()

如图

计算样本数据之前，需要对特征值做归一化

def autoNorm (dataSet):
    """
           归一化数据
    """
    minVals = dataSet.min(0)
    maxVals = dataSet.max(0)
    ranges = maxVals - minVals
    normDataSet = np.zeros(np.shape(dataSet))
    m = dataSet.shape[0]                     #1000 读取矩阵长度
    normDataSet = dataSet - np.tile(minVals, (m,1)) #tile 生成最小值矩阵 根据训练数据大小生成 矩阵减法 减去最小值
    #print(np.tile(minVals, (m,1)))
    normDataSet = normDataSet/np.tile(ranges, (m,1)) #除以（最大值-最小值）
    return normDataSet, ranges, minVals

测试代码

def datingClassTest(filname):
    hoRatio = 0.10
    datingDateMat, datingLabels = knn.file2maxtrix(filname)#读取训练数据
    normMat, ranges, minVals = knn.autoNorm(datingDateMat)#归一化训练数据
    m = normMat.shape[0] #返回矩阵行数
    numTestVecs = int (m*hoRatio)
    errorCount = 0.0
    for i in range(numTestVecs): 
        #前100作为验证样本 后900作为训练样本
        classifierResult = classify0(normMat[i, :], normMat[numTestVecs:m, :], datingLabels[numTestVecs:m],3)
        print("the classifier came back with:%d, the real answer is : %d" % (classifierResult, datingLabels[i]))
        
        if classifierResult != datingLabels[i] :#计算错误率
            errorCount += 1.0
    
    print("the total error rate is:%f"%(errorCount/float(numTestVecs)))
    return

输出：

the classifier came back with:3, the real answer is : 3
the classifier came back with:2, the real answer is : 2
the classifier came back with:1, the real answer is : 1
the classifier came back with:1, the real answer is : 1
the classifier came back with:1, the real answer is : 1
the classifier came back with:1, the real answer is : 1
……
the classifier came back with:3, the real answer is : 1
the total error rate is:0.050000

100个样本错了5个 错误率为0.05

根据用户输入返回分类结果

def classifyPerson(filname):
    resultList = ['not at all', 'in small doses', 'in large doses']
    percentTats = float(input('percentage of time spent playing video games?'))
    ffMiles = float(input('frequent flier miles earned per year?'))
    iceCream = float(input('liters of ice cream consumed per year?'))
    
    datingDateMat, datingLabels = knn.file2maxtrix(filname)#读取训练数据
    normMat, ranges, minVals = knn.autoNorm(datingDateMat)#归一化训练数据
    
    inArr = np.array([percentTats, ffMiles, iceCream])
    inArrNorm = inArr-minVals/ranges #对测试数据归一化
    classifierResult = knn.classify0(inArrNorm,normMat,datingLabels,3)
    
    print("you will probably like this person:",resultList[classifierResult-1])
    
    return

输入输出内容

percentage of time spent playing video games?10
frequent flier miles earned per year?10000
liters of ice cream consumed per year?0.5
you will probably like this person: not at all