机器学习实战python3——k-近邻算法

kNN算法伪代码

对未知类别属性的数据集中的每个点依次执行以下操作：

1. 计算已知类别数据中的每个点与当前点之间的距离；
2. 按照距离递增次序排序；
3. 选取与当前点距离最小的k个点；
4. 确定前k个点所在类别的出现频率；
5. 返回前k个点出现频率最高的类别作为当前点的预测分类。

from numpy import *
from os import listdir
import operator

# knn算法
def classify0(inX, dataSet, labels, k):
    dataSetSize = dataSet.shape[0]
    # 计算距离
    diffMat = tile(inX, (dataSetSize, 1)) - dataSet
    sqDiffMat = pow(diffMat, 2)
    sqDistance = sqDiffMat.sum(axis=1)
    distances = pow(sqDistance, 0.5)
    # 选择距离最小的k个点
    sortedDistance = distances.argsort()  # 对距离进行排序，返回排序后的元素下标
    class_count = {}
    for i in range(k):
        vote_ilabels = labels[sortedDistance[i]]
        # 返回键“vote_ilabels”的值，不存在返回0
        class_count[vote_ilabels] = class_count.get(vote_ilabels, 0) + 1 # 专门对列表计数
    # 排序
    sorted_classcount = sorted(class_count.items(),
                               key = operator.itemgetter(1) , reverse = True)  # 按照值排序，降序
    return sorted_classcount[0][0]


# 2-2文本转化为numpy解析
def file2matrix(filename):
    fr = open(filename)
    array_olines = fr.readlines() # 所有行读成列表
    # 获得文件行数
    lines_counts = len(array_olines)
    # 创建一个全为0的矩阵
    return_mat = zeros((lines_counts, 3))
    class_lable_vector = []
    index = 0
    # 解析文件数据到列表
    for line in array_olines:
        line = line.strip()
        list_from_line = line.split('\t')
        return_mat[index, : ] = list_from_line[0:3]
        class_lable_vector.append((int(list_from_line[-1])))
        index += 1
    return return_mat, class_lable_vector


# 2-3准备数据，归一化数值
def auto_norm(dataSet):
    minVals = dataSet.min(0) # 每一列的最小值
    maxVals = dataSet.max(0) # 每一列的最大值
    ranges = maxVals - minVals #每一列中最小最大值的差值
    normDataSet = zeros(shape(dataSet)) #创建一个和dataset维度一样的全为0矩阵
    m = dataSet.shape[0] # dataSet的行数
    normDataSet = dataSet - tile(minVals, (m, 1)) # oldValue-min
    normDataSet = normDataSet / tile(ranges, (m, 1)) # （oldValue-min）/ (max -min)
    return normDataSet, ranges, minVals


# 2-4 测试算法
def datingClassTest():
    hoRatio = 0.1
    datingDataMat,datingLables = file2matrix('data_test_set2.txt')
    normMat,ranges, minVals = auto_norm(datingDataMat)
    m = normMat.shape[0]
    numTsetVecs = int(m * hoRatio)
    errorCount = 0.0
    for i in range (numTsetVecs):
        classifierResult = classify0(normMat[i, :],
                                     normMat[numTsetVecs:m, :],
                                     datingLables[numTsetVecs:m],
                                     4)
        print("The classifier come back with: %d, The real answer is: %d" \
              %(classifierResult, datingLables[i]))
        if classifierResult != datingLables[i]:
            errorCount += 1.0
    print("The total error rate is :%.4f" %(errorCount / float(numTsetVecs)))

datingClassTest()
print(file2matrix('data_test_set2.txt')[0].shape)


# 2-5 约会网站预测函数
def classifyPerson():
    resultList = ['not at al', 'in small doses', 'in large doses']
    # input函数允许用户自己输入数值
    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?"))
    datingDataMat, datingLabels = file2matrix('data_test_set2.txt')
    normMat, ranges, minVals = auto_norm(datingDataMat)
    inArr = array([ffMiles, percentTats, iceCream])
    classifyResult = classify0((inArr - minVals)/ ranges,
                               normMat,
                               datingLabels,
                               3)

    print("You will probably like this person: ", resultList[classifyResult - 1])
#classifyPerson()


# 2.3 手写识别系统
def img2vect(filename):
    returnVect = zeros((1, 1024))
    fr = open(filename)
    for i in range(32):
        lineStr = fr.readline() # 读取一行
        for j in range(32):
            returnVect[0, 32 * i + j] = int(lineStr[j])
    return returnVect
# print(img2vect('trainingDigits/0_2.txt')[0, 0:32])


# 2-6 手写数字识别系统测试算法
def handwriteingClassTest():
    # 获取目录内容
    hwLables = []
    trainFileList = listdir('trainingDigits') # 目录下所有文件名生产一个列表
    m = len(trainFileList)
    trainMat = zeros((m, 1024))
    # 从文件名中解析分类数字
    for i in range(m):
        fileNameStr = trainFileList[i]
        fileStr = fileNameStr.split('.')[0]
        classNumstr = int(fileStr.split('_')[0])
        hwLables.append(classNumstr)
        trainMat[i, :] = img2vect('trainingDigits/%s' % fileNameStr)
    listFileList = listdir('testDigits')
    errorCount = 0.0
    mTest = len(listFileList)
    for i in range(mTest):
        fileNameStr = listFileList[i]
        fileStr = fileNameStr.split(',')[0]
        classNumstr = int(fileStr.split('_')[0])
        vectorUnderTest = img2vect('testDigits/%s' % fileNameStr)
        classifierResult = classify0(vectorUnderTest,
                                     trainMat,
                                     hwLables,
                                     3)
        print('the classifier came back with :%d, the real answer is :%d'\
              %(classifierResult, classNumstr))
        if classifierResult != classNumstr:
            errorCount += 1.0
    print('the total of errors is: %d' %(errorCount))
    print('the total error rate is : %f' %(errorCount / mTest) )
    print(m,mTest)
handwriteingClassTest()

代码运行结果如下