1.4 KNN算法学习——KNN算法的实现与封装

封装代码实现

import numpy as np
from math import sqrt
from collections import Counter

def kNN_classfy(k,X_train,y_train,x):

    assert 1 <= k <=X_train.shape[0],"k must be valid"
    assert X_train.shape[0]==y_train.shape[0],"the size of X_train must equal to the sise of y_train"
    assert X_train.shape[1]==x.shape[0],"the feature number of x must be equal to X_train"

    #距离求解
    distance = [sqrt(np.sum((x_train-x)**2)) for x_train in X_train]

    #最近距离的索引排序
    nearest=np.argsort(distance)

    #求前k个最近的点的y值
    top_K=[y_train[i] for i in nearest[:k]]

    #投票
    votes=Counter(top_K)

    #返回投票结果
    return votes.most_common(1)[0][0]

运行结果：

结果相同（%run 文件名运行脚本文件）

模仿sklearn中的KNN模型改进封装代码

import numpy as np
from math import sqrt
from collections import Counter

def kNN_classfy(k,X_train,y_train,x):

    assert 1 <= k <=X_train.shape[0],"k must be valid"
    assert X_train.shape[0]==y_train.shape[0],"the size of X_train must equal to the sise of y_train"
    assert X_train.shape[1]==x.shape[0],"the feature number of x must be equal to X_train"

    #距离求解
    distance = [sqrt(np.sum((x_train-x)**2)) for x_train in X_train]

    #最近距离的索引排序
    nearest=np.argsort(distance)

    #求前k个最近的点的y值
    top_K=[y_train[i] for i in nearest[:k]]

    #投票
    votes=Counter(top_K)

    #返回投票结果
    return votes.most_common(1)[0][0]

运行结果：

注：KNN模型测试集要求输入的矩阵，所以需要将x进行reshape变换为矩阵