KNN 手写数字识别

import operator
from numpy import *
import os

'''
    需要文件 可以给我发消息
    图像形状 32x32
    改成 1x32x32的样式
'''

def classify0(inX, dateSet, labels, k):
    dateSetSize = dateSet.shape[0]
    diffMat = tile(inX, (dateSetSize, 1)) - dateSet #差
    sqDiffMat = diffMat**2#平方
    sqDistances = sqDiffMat.sum(axis = 1)# 求和
    distances = sqDistances**0.5 #开方
    sortedDistIndicies = distances.argsort()  #排序
    classCount = {}
    for i in range(k):
        voteILabel = labels[sortedDistIndicies[i]] #标签
        classCount[voteILabel] = classCount.get(voteILabel, 0) + 1
    sortedClassCount = sorted(classCount.items(), key = operator.itemgetter(1), reverse = True)
    return sortedClassCount[0][0]
        


def img2vector(filename):
    returnVect = zeros((1, 1024))  #1x32x32
    fr = open(filename)# 打开文件
    for i in range(32):# 32行
        lineStr = fr.readline()# 读取一行
        for j in range(32):
            returnVect[0, 32*i+j] =int(lineStr[j])# 换算成一行
    return returnVect

def handwritingClassTest():
    hwLabels = []
    traingingFileList = os.listdir('C:\\Users\\lvcun\\Desktop\\trainingDigits') #获取训练集
    m = len(traingingFileList)#有多少个文件名字
    traingingMat = zeros((m, 1024))# 每一个数字都改成1x1024的样子  就有mx1x1024
    for i in range(m):#对于每一个文件
        fileNameStr = traingingFileList[i]#当前文件名字
        fileStr = fileNameStr.split('.')[0]#分割   (0_1.txt) --> 0_1
        classNumStr = int(fileStr.split('_')[0])  #分割  (0_1) --> 0
        hwLabels.append(classNumStr) #当前数字标签
        traingingMat[i, :]=img2vector('C:\\Users\\lvcun\\Desktop\\trainingDigits\\%s'%fileNameStr) # 将当前的数字32x32 转换成1x1024
    testFileList = os.listdir('C:\\Users\\lvcun\\Desktop\\testDigits') #获取测试集
    errorCount = 0.0#错误率
    mTest = len(testFileList)#多少个训练数字
    for i in range(mTest):#对于每个训练数字
        fileNameStr = testFileList[i] #获取当前文件名
        fileStr = fileNameStr.split('.')[0]#分割 (0_1.txt)  --> 0_1
        classNumStr = int(fileStr.split('_')[0])#分割(0_1)  --> 0
        vvectorUnderTest = img2vector('C:\\Users\\lvcun\\Desktop\\testDigits\\%s'%fileNameStr) # 将当前的数字32x32 转换成1x1024
        classifierResult = classify0(vvectorUnderTest, traingingMat, hwLabels, 3)#计算距离
        print("the classifier came back with: %d, the real answer is: %d"%(classifierResult, classNumStr))
        if classifierResult != classNumStr :
            errorCount += 1
    print('\nthe total number of errors is:%d'%errorCount)
    print('\nthe total error rate is: %f'%(errorCount/float(mTest)))

handwritingClassTest()