《机器学习实战》之k-近邻算法（手写识别系统）

这个玩意和改进约会网站的那个差不多，它是提前把所有数字转换成了32*32像素大小的黑白图，然后转换成字符图（用0,1表示），将所有1024个像素点用一维矩阵保存下来，这样就可以通过knn计算欧几里得距离来得到最接近的答案。

import os
import operator
from numpy import *

def classify0(inX, dataSet, labels, k):
dataSetSize = dataSet.shape[0]
diffMat = tile(inX, (dataSetSize,1)) - dataSet  #统一矩阵，实现加减
sqDiffMat = diffMat**2
sqDistances = sqDiffMat.sum(axis=1)  #进行累加，axis=0是按列，axis=1是按行
distances = sqDistances**0.5  #开根号
sortedDistIndicies = distances.argsort()  #按升序进行排序，返回原下标
classCount = {}
for i in range(k):
voteIlabel = labels[sortedDistIndicies[i]]
classCount[voteIlabel] = classCount.get(voteIlabel, 0) + 1  #get是字典中的方法，前面是要获得的值，后面是若该值不存在时的默认值
sortedClassCount = sorted(classCount.items(), key=operator.itemgetter(1), reverse=True)
return sortedClassCount[0][0]

def img2vector(filename):
f = open(filename)
returnVect = zeros((1,1024))
for i in range(32):
line = f.readline()
for j in range(32):
returnVect[0,i*32+j] = int(line[j])
return returnVect

def handwritingClassTest():
fileList = os.listdir('trainingDigits')
m = len(fileList)
traingMat = zeros((m, 1024))
hwlabels = []
for i in range(m):
fileName = fileList[i]
prefix = fileName.split('.')[0]
number = int(prefix.split('_')[0])
hwlabels.append(number)
traingMat[i,:] = img2vector('trainingDigits/%s' %fileName)
testFileList = os.listdir('testDigits')
m = len(testFileList)
errorNum = 0.0
for i in range(m):
testFileName = testFileList[i]
prefix = testFileList[i].split('.')[0]
realNumber = int(prefix.split('_')[0])
testMat = img2vector('testDigits/%s' %testFileName)
testResult = classify0(testMat, traingMat, hwlabels, 3)
if testResult != realNumber:
errorNum += 1
print('The classifier came back with: %d, the real answer is: %d' %(testResult, realNumber))
print('错误率为%f' %(errorNum/float(m)))

if __name__ == '__main__':
handwritingClassTest()