kNN算法

纸上得来终觉浅，觉知此事要躬行，看《机器学习实战》这本书的时候，只是看理解不了代码，自己动手敲一遍，一行一行地理解代码才能弄明白每一句代码的意思。

from numpy import *
import operator

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


group, labels = createDateSet()
# print(group)
'''
4行1列
[[1.  1.1]
 [1.  1. ]
 [0.  0. ]
 [0.  0.1]]
'''
# print(labels)


def classify0(inX, dataSet, labels, k):
    print(dataSet)  # group
    print(type(dataSet))  # <class 'numpy.ndarray'>
    dataSetSize = dataSet.shape[0]
    # print(dataSetSize)  # 4
    diffMat = tile(inX, (dataSetSize, 1)) - dataSet  # 将inX复制4行1列
    print(diffMat)
    '''
    [[0.  0. ]
     [0.  0.1]
     [1.  1.1]
     [1.  1. ]]
    '''
    sqDiffMat = diffMat**2  # 对每个元素都取平方
    print(sqDiffMat)
    '''
    [[0.   0.  ]
     [0.   0.01]
     [1.   1.21]
     [1.   1.  ]]
    '''
    print('---------------------')
    sqDistances = sqDiffMat.sum(axis=1)
    print(sqDistances)
    '''[0.   0.01 2.21 2.  ]'''
    distances = sqDistances**0.5
    print(distances)  # [0.         0.1        1.48660687 1.41421356]
    sortedDistIndicies = distances.argsort()  # 按距离从小到大排列
    print(sortedDistIndicies)  # [0 1 3 2]
    classCount = {}
    for i in range(k):
        voteIlabel = labels[sortedDistIndicies[i]]
        classCount[voteIlabel] = classCount.get(voteIlabel, 0) + 1
    sortedClassCount = sorted(classCount.items(), key=operator.itemgetter(1), reverse=True)
    return sortedClassCount[0][0]


print(classify0([1, 1.1], group, labels, 3))  # A

https://blog.csdn.net/lionprinceton/article/details/18620809
NumPy矩阵与数组的区别：矩阵只能是2维，矩阵的行与行之间用分号隔开
https://blog.csdn.net/qq_28481231/article/details/76619040