原文地址:http://www.alannah.cn/2019/04/03/numpye/
计算两个矩阵行向量之间的欧氏距离
有时我们需要计算两个矩阵行向量之间的欧氏距离,但是在numpy中,没有函数可以直接进行两个矩阵行向量之间的距离计算。
假设:x(m,n),y(k,n),计算xy两个矩阵行向量之间的欧氏距离,则dis(m,k)
将问题转化为(x-y)^2 ==> x^2 + y^2 -2xy
计算代码如下:
def EuclideanDistance(x, y):
"""
get the Euclidean Distance between to matrix
(x-y)^2 = x^2 + y^2 - 2xy
:param x:
:param y:
:return:
"""
(rowx, colx) = x.shape
(rowy, coly) = y.shape
if colx != coly:
raise RuntimeError('colx must be equal with coly')
xy = np.dot(x, y.T)
x2 = np.repeat(np.reshape(np.sum(np.multiply(x, x), axis=1), (rowx, 1)), repeats=rowy, axis=1)
y2 = np.repeat(np.reshape(np.sum(np.multiply(y, y), axis=1), (rowy, 1)), repeats=rowx, axis=1).T
dis = x2 + y2 - 2 * xy
return dis