kNN算法python实现和简单数字识别的方法

def createDataSet():
    group = array([[1.0, 0.9], [1.0, 1.0], [0.1, 0.2], [0.0, 0.1]])
    labels = ['A','A','B','B']
    return group,labels
#inputX表示输入向量（也就是我们要判断它属于哪一类的）
#dataSet表示训练样本
#label表示训练样本的标签
#k是最近邻的参数，选最近k个
def kNNclassify(inputX, dataSet, labels, k):
    dataSetSize = dataSet.shape[0]#计算有几个训练数据
    #开始计算欧几里得距离
    diffMat = tile(inputX, (dataSetSize,1)) - dataSet
   
    sqDiffMat = diffMat ** 2
    sqDistances = sqDiffMat.sum(axis=1)#矩阵每一行向量相加
    distances = sqDistances ** 0.5
    #欧几里得距离计算完毕
    sortedDistance = distances.argsort()
    classCount = {}
    for i in xrange(k):
        voteLabel = labels[sortedDistance[i]]
        classCount[voteLabel] = classCount.get(voteLabel,0) + 1
    res = max(classCount)
    return res

def main():
    group,labels = createDataSet()
    t = kNNclassify([0,0],group,labels,3)
    print t
   
if __name__=='__main__':
    main()

def createDataSet():
    group = array([[1.0, 0.9], [1.0, 1.0], [0.1, 0.2], [0.0, 0.1]])
    labels = ['A','A','B','B']
    return group,labels
#inputX表示输入向量（也就是我们要判断它属于哪一类的）
#dataSet表示训练样本
#label表示训练样本的标签
#k是最近邻的参数，选最近k个
def kNNclassify(inputX, dataSet, labels, k):
    dataSetSize = dataSet.shape[0]#计算有几个训练数据
    #开始计算欧几里得距离
    diffMat = tile(inputX, (dataSetSize,1)) - dataSet
    #diffMat = inputX.repeat(dataSetSize, aixs=1) - dataSet
    sqDiffMat = diffMat ** 2
    sqDistances = sqDiffMat.sum(axis=1)#矩阵每一行向量相加
    distances = sqDistances ** 0.5
    #欧几里得距离计算完毕
    sortedDistance = distances.argsort()
    classCount = {}
    for i in xrange(k):
        voteLabel = labels[sortedDistance[i]]
        classCount[voteLabel] = classCount.get(voteLabel,0) + 1
    res = max(classCount)
    return res

def img2vec(filename):
    returnVec = zeros((1,1024))
    fr = open(filename)
    for i in range(32):
        lineStr = fr.readline()
        for j in range(32):
            returnVec[0,32*i+j] = int(lineStr[j])
    return returnVec
   
def handwritingClassTest(trainingFloder,testFloder,K):
    hwLabels = []
    trainingFileList = os.listdir(trainingFloder)
    m = len(trainingFileList)
    trainingMat = zeros((m,1024))
    for i in range(m):
        fileName = trainingFileList[i]
        fileStr = fileName.split('.')[0]
        classNumStr = int(fileStr.split('_')[0])
        hwLabels.append(classNumStr)
        trainingMat[i,:] = img2vec(trainingFloder+'/'+fileName)
    testFileList = os.listdir(testFloder)
    errorCount = 0.0
    mTest = len(testFileList)
    for i in range(mTest):
        fileName = testFileList[i]
        fileStr = fileName.split('.')[0]
        classNumStr = int(fileStr.split('_')[0])
        vectorUnderTest = img2vec(testFloder+'/'+fileName)
        classifierResult = kNNclassify(vectorUnderTest, trainingMat, hwLabels, K)
        #print classifierResult,' ',classNumStr
        if classifierResult != classNumStr:
            errorCount +=1
    print 'tatal error ',errorCount
    print 'error rate',errorCount/mTest
       
def main():
    t1 = time.clock()
    handwritingClassTest('trainingDigits','testDigits',3)
    t2 = time.clock()
    print 'execute ',t2-t1
if __name__=='__main__':
    main()