Numpy实现两个矩阵欧式距离最小下标返回

最新推荐文章于 2024-06-15 08:00:00 发布

yangdelu855

最新推荐文章于 2024-06-15 08:00:00 发布

阅读量236

点赞数

分类专栏： numpy 文章标签： numpy 矩阵 python

本文链接：https://blog.csdn.net/yangdelu855/article/details/129198376

版权

numpy 专栏收录该内容

1 篇文章 0 订阅

订阅专栏

需要注意两个矩阵维度不同的时候，不可以直接做减运算，需要对某个矩阵扩增维度，然后求和降维

# -*- encoding: utf-8 -*-

import tensorflow as tf
import torch
import numpy as np
import math
def cal_dist(a,b):
    ans = 0
    print(a,b)
    for x,y in zip(a,b):
        ans += (x-y)**2
    return math.sqrt(ans)
def get_distance(x,y):
    dist = []
    for data_x in x:
        tmp = []
        for data_y in y:
            tmp.append(cal_dist(data_x,data_y))
        dist.append(tmp)
    return dist
if __name__ == '__main__':
    m, n, d = 3, 2, 2
    x = np.array([[1, 2], [3, 4], [5, 6]])
    y = np.array([[2, 3], [4, 5]])
    x_ = np.repeat(x,n,0)
    x_ = x_.reshape([m,n,d])
    print(x_-y)
    dist_np = np.sqrt(np.sum(np.square((x_-y)),axis=-1))
    dist_np2 = np.linalg.norm(x_ - y,axis=-1,ord=2)
    dist = get_distance(x,y)
    print(dist_np)
    print(dist_np2)
    print(dist)
    print(np.argmin(dist_np,axis=-1))
    print(np.argmin(dist_np2,axis=-1))
    print(np.argmin(dist_np,axis=-1))

#sprt(((x1-y1)**2+(x2-y2)**2))