from numpy import *
import operator

def createDataSet() :
    group = array([[1.0, 1.1], [1.0, 1.0], [0, 0], [0, 1.1]])
    labels = [‘A‘, ‘A‘, ‘B‘, ‘B‘]
    return group, labels


‘‘‘
    tile(array, (intR, intC): 對矩陣進行組合，縱向復制intR次， 橫向復制intC次
    比如 : tile([1,2,3], (3, 2))
    輸出
    [
        [1, 2, 3, 1, 2, 3],
        [1, 2, 3, 1, 2, 3],
        [1, 2, 3, 1, 2, 3]
    ]
    array減法， 兩個 行列數相等的矩陣，對應位置做減法

    argsort(array, axis) 對矩陣進行排序，axis=0 按列排序， axis=1 按行排序  輸出的是排序的索引。比如輸出[0,2,1], 排序結果結果為 array[0],array[2].array[1]

    aorted(iteratorItems, key, reverse)  對可叠代的對象進行排序

 
‘‘‘
def classify0(intX, dataSet, labels, k) :    # 假設輸入 intX = [0, 0], dataSet = array([[1.0, 1.1], [1.0, 1.0], [0, 0], [0, 1.1]]), labels = [‘A‘, ‘A‘, ‘B‘, ‘B‘], k = 3
    dataSetSize = dataSet.shape[0];    # 行數 dataSetSize = 4
    diffMat = tile(intX, (dataSetSize, 1)) - dataSet    # 矩陣差  diffMat = array([[-1.0, -1.1], [-1.0, -1.0], [0, 0], [0, -1.1]])
 

    sqDiffMat = diffMat ** 2    # 平方   sqDiffMat = array([[1, 1.21], [1, 1], [0, 0], [0, 1.21]])
    sqDistances = sqDiffMat.sum(axis=1)    # 行和，axis=0時輸出縱和  sqDistances = array([2.21, 2, 0, 1.21])
    distances = sqDistances ** 0.5    # 開平方 distances = array([1.41, 1.48, 0, 1.1])
    sortedDistIndicies = distances.argsort()    #
 
 排序 sortedDistIndicies = array([2, 3, 0, 1])
    classCount = {}
    for i in range(k) :
        voteIlabel = labels[sortedDistIndicies[i]]   
        classCount[voteIlabel] = classCount.get(voteIlabel, 0) + 1    # {label:count}， 取距離最小的三個， 統計label出現的次數，最終 classCount = {‘B‘: 2, ‘A‘: 1}
    sortedClassCount = sorted(classCount.items(), key=operator.itemgetter(1), reverse=True)    # 對dict按照v值（distance）進行倒序排序 sortedClassCount = [(‘B‘, 2), (‘A‘, 1)]
    return sortedClassCount[0][0]  # 返回第一個tuple的第一個值，也就是出現次數最高的label， 這裏返回‘B’


if __name__ == "__main__":
    group, labels = createDataSet()
    classify0([0, 0], group, labels, 3)

[Machine :Learning] kNN近鄰算法