k-nearest neighbors[k-邻近法]

import numpy as np
import operator as op

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


group,labels = createDataSet()
# print(group, labels)

'''
inX:分类的输入向量
dataSet:输入的训练样本集
labels:标签向量
k：选择最近邻的数目

欧式距离
'''
def classify(inX, dataSet, labels, k):
    dataSetSize = dataSet.shape[0]
    # tile 复制矩阵dataSetSize行，然后矩阵相减
    diffMat = np.tile(inX, (dataSetSize,1)) - dataSet
    sqDifMat = diffMat**2
    # 矩阵每列的和相加，保存到第一个
    sqDistance = sqDifMat.sum(axis = 1)
    distance = sqDistance**0.5
    #排序返回索引
    sortedDistIndicies = distance.argsort()
    classCount = {}
    for i in range(k):

        voteIlabel = labels[sortedDistIndicies[i]]
        # 字典通过key返回value；且一定要加0
        classCount[voteIlabel] = classCount.get(voteIlabel, 0) + 1


    # 通过公共建对字典序列进行排序
    sortedClassCount = sorted(classCount.items(), key=op.itemgetter(1),reverse=True)
    return sortedClassCount[0][0]

ret = classify([1,1], group, labels, 3)
print(ret)

-->A