统计学习方法3-python实现KNN线性扫描算法

1. 使用python 自实现

import numpy as np

def o_distance(x1,y1):
    return np.sqrt(np.sum(np.square(x1-y1)))

class KNN:
    def __init__(self,X,y):
        self.X = X
        self.y = y
        
    def classify(self,x):
        distance = [o_distance(x,i) for i in self.X]
        print('distance:',distance)
        min_dist = np.argmin(distance)
        print('near point:',X[min_dist])
        return y[min_dist]
        
if __name__ == "__main__":
    X = np.array([(5,4),(9,6),(4,7),(2,3),(8,1),(7,2)])
    y = np.array([1,1,1,-1,-1,-1])
    point = np.array((5,3))
    knn = KNN(X,y)
    print('point class: ',knn.classify(point))

2. python 调库实现

import numpy as np
from sklearn.neighbors import KNeighborsClassifier

if __name__ == "__main__":
    X = np.array([(5,4),(9,6),(4,7),(2,3),(8,1),(7,2)])
    y = np.array([1,1,1,-1,-1,-1])
    point = np.array([(5,3)])
    
    for i in range(len(X)):
        k = i + 1 
        knn = KNeighborsClassifier(n_neighbors= k)
        knn.fit(X,y)
        k_class = knn.predict(point)
        print('k = {}, point class: {}'.format(k,k_class))