《机器学习实战》之k-近邻算法（手写识别系统）

Posted 2020-10-24 谦谦君子，陌上其华

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。

 1 import os
 2 import operator
 3 from numpy import *
 4 
 5 def classify0(inX, dataSet, labels, k):
 6     dataSetSize = dataSet.shape[0]
 7     diffMat = tile(inX, (dataSetSize,1)) - dataSet  #统一矩阵，实现加减
 8     sqDiffMat = diffMat**2
 9     sqDistances = sqDiffMat.sum(axis=1)  #进行累加，axis=0是按列，axis=1是按行
10     distances = sqDistances**0.5  #开根号
11     sortedDistIndicies = distances.argsort()  #按升序进行排序，返回原下标
12     classCount = {}
13     for i in range(k):
14         voteIlabel = labels[sortedDistIndicies[i]]
15         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
16     sortedClassCount = sorted(classCount.items(), key=operator.itemgetter(1), reverse=True)
17     return sortedClassCount[0][0]
18 
19 
20 def img2vector(filename):
21     f = open(filename)
22     returnVect = zeros((1,1024))
23     for i in range(32):
24         line = f.readline()
25         for j in range(32):
26             returnVect[0,i*32+j] = int(line[j])
27     return returnVect
28 
29 
30 def handwritingClassTest():
31     fileList = os.listdir(\'trainingDigits\')
32     m = len(fileList)
33     traingMat = zeros((m, 1024))
34     hwlabels = []
35     for i in range(m):
36         fileName = fileList[i]
37         prefix = fileName.split(\'.\')[0]
38         number = int(prefix.split(\'_\')[0])
39         hwlabels.append(number)
40         traingMat[i,:] = img2vector(\'trainingDigits/%s\' %fileName)
41     testFileList = os.listdir(\'testDigits\')
42     m = len(testFileList)
43     errorNum = 0.0
44     for i in range(m):
45         testFileName = testFileList[i]
46         prefix = testFileList[i].split(\'.\')[0]
47         realNumber = int(prefix.split(\'_\')[0])
48         testMat = img2vector(\'testDigits/%s\' %testFileName)
49         testResult = classify0(testMat, traingMat, hwlabels, 3)
50         if testResult != realNumber:
51             errorNum += 1
52         print(\'The classifier came back with: %d, the real answer is: %d\' %(testResult, realNumber))
53     print(\'错误率为%f\' %(errorNum/float(m)))
54 
55 if __name__ == \'__main__\':
56     handwritingClassTest()