K近鄰算法及其Python實現

下面貼出Python代碼

knnClassify.py

from numpy import *
import operator

def creatDataSet():
    group = array([[1.0,1.1],[1.0,1.0],[0,0],[0,0.1]])
    labels = ['A','A','B','B']
    return group, labels

def classify(inX,dataSet,labels,k):
    numSamples = dataSet.shape[0]
    diffMat = tile(inX,(numSamples,1)) - dataSet
    sqDiffMat = diffMat**2
    sqDistances = sqDiffMat.sum(axis = 1)
    distances = sqDistances ** 0.5
    sortedDistIndicies = distances.argsort()
    classCount = {}
    for i in xrange(k):
        voteILabel = labels[sortedDistIndicies[i]]
        classCount[voteILabel] = classCount.get(voteILabel,0) + 1
    maxLabel = sorted(classCount.iteritems(),
                      key = operator.itemgetter(1),reverse = True)
    return maxLabel[0][0]

if __name__=="__main__":
    g,l = creatDataSet()
    labelOfinX = classify([1.0,1.2],g,l,1)
    print labelOfinX