kNN Code?

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

Attached please find the kNN Python code from "Machine Learning in Action by Peter Harrington". Unfortunately it is not running. Anybody can take a look and tell me how to fix it(I use Spyder(Python 2.7)?
Thanks

1. '''
   
2. Created on Sep 16, 2010
   
3. kNN: k Nearest Neighbors
   
4. 
   
5. Input:      inX: vector to compare to existing dataset (1xN)
   
6.             dataSet: size m data set of known vectors (NxM)
   
7.             labels: data set labels (1xM vector)
   
8.             k: number of neighbors to use for comparison (should be an odd number)
   
9.             
   
10. Output:     the most popular class label
    
11. 
    
12. @author: pbharrin
    
13. '''
    
14. from numpy import *
    
15. import operator
    
16. from os import listdir
    
17. 
    
18. def classify0(inX, dataSet, labels, k):
    
19.     dataSetSize = dataSet.shape[0]
    
20.     diffMat = tile(inX, (dataSetSize,1)) - dataSet
    
21.     sqDiffMat = diffMat**2
    
22.     sqDistances = sqDiffMat.sum(axis=1)
    
23.     distances = sqDistances**0.5
    
24.     sortedDistIndicies = distances.argsort()     
    
25.     classCount={}         
    
26.     for i in range(k):
    
27.         voteIlabel = labels[sortedDistIndicies[i]]
    
28.         classCount[voteIlabel] = classCount.get(voteIlabel,0) + 1
    
29.     sortedClassCount = sorted(classCount.iteritems(), key=operator.itemgetter(1), reverse=True)
    
30.     return sortedClassCount[0][0]
    
31. 
    
32. def createDataSet():
    
33.     group = array([[1.0,1.1],[1.0,1.0],[0,0],[0,0.1]])
    
34.     labels = ['A','A','B','B']
    
35.     return group, labels
    
36. 
    
37. def file2matrix(filename):
    
38.     fr = open(filename)
    
39.     numberOfLines = len(fr.readlines())         #get the number of lines in the file
    
40.     returnMat = zeros((numberOfLines,3))        #prepare matrix to return
    
41.     classLabelVector = []                       #prepare labels return   
    
42.     fr = open(filename)
    
43.     index = 0
    
44.     for line in fr.readlines():
    
45.         line = line.strip()
    
46.         listFromLine = line.split('\t')
    
47.         returnMat[index,:] = listFromLine[0:3]
    
48.         classLabelVector.append(int(listFromLine[-1]))
    
49.         index += 1
    
50.     return returnMat,classLabelVector
    
51.    
    
52. def autoNorm(dataSet):
    
53.     minVals = dataSet.min(0)
    
54.     maxVals = dataSet.max(0)
    
55.     ranges = maxVals - minVals
    
56.     normDataSet = zeros(shape(dataSet))
    
57.     m = dataSet.shape[0]
    
58.     normDataSet = dataSet - tile(minVals, (m,1))
    
59.     normDataSet = normDataSet/tile(ranges, (m,1))   #element wise divide
    
60.     return normDataSet, ranges, minVals
    
61.    
    
62. def datingClassTest():
    
63.     hoRatio = 0.50      #hold out 10%
    
64.     datingDataMat,datingLabels = file2matrix('datingTestSet2.txt')       #load data setfrom file
    
65.     normMat, ranges, minVals = autoNorm(datingDataMat)
    
66.     m = normMat.shape[0]
    
67.     numTestVecs = int(m*hoRatio)
    
68.     errorCount = 0.0
    
69.     for i in range(numTestVecs):
    
70.         classifierResult = classify0(normMat[i,:],normMat[numTestVecs:m,:],datingLabels[numTestVecs:m],3)
    
71.         print "the classifier came back with: %d, the real answer is: %d" % (classifierResult, datingLabels[i])
    
72.         if (classifierResult != datingLabels[i]): errorCount += 1.0
    
73.     print "the total error rate is: %f" % (errorCount/float(numTestVecs))
    
74.     print errorCount
    
75.    
    
76. def img2vector(filename):
    
77.     returnVect = zeros((1,1024))
    
78.     fr = open(filename)
    
79.     for i in range(32):
    
80.         lineStr = fr.readline()
    
81.         for j in range(32):
    
82.             returnVect[0,32*i+j] = int(lineStr[j])
    
83.     return returnVect
    
84. 
    
85. def handwritingClassTest():
    
86.     hwLabels = []
    
87.     trainingFileList = listdir('trainingDigits')           #load the training set
    
88.     m = len(trainingFileList)
    
89.     trainingMat = zeros((m,1024))
    
90.     for i in range(m):
    
91.         fileNameStr = trainingFileList[i]
    
92.         fileStr = fileNameStr.split('.')[0]     #take off .txt
    
93.         classNumStr = int(fileStr.split('_')[0])
    
94.         hwLabels.append(classNumStr)
    
95.         trainingMat[i,:] = img2vector('trainingDigits/%s' % fileNameStr)
    
96.     testFileList = listdir('testDigits')        #iterate through the test set
    
97.     errorCount = 0.0
    
98.     mTest = len(testFileList)
    
99.     for i in range(mTest):
    
100.         fileNameStr = testFileList[i]
     
101.         fileStr = fileNameStr.split('.')[0]     #take off .txt
     
102.         classNumStr = int(fileStr.split('_')[0])
     
103.         vectorUnderTest = img2vector('testDigits/%s' % fileNameStr)
     
104.         classifierResult = classify0(vectorUnderTest, trainingMat, hwLabels, 3)
     
105.         print "the classifier came back with: %d, the real answer is: %d" % (classifierResult, classNumStr)
     
106.         if (classifierResult != classNumStr): errorCount += 1.0
     
107.     print "\nthe total number of errors is: %d" % errorCount
     
108.     print "\nthe total error rate is: %f" % (errorCount/float(mTest))

复制代码

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏0 回帖

关键词：code COD knn ODE handwriting comparison existing running compare please