numpy矩阵乘法每个col和sum

Question

我有一些矩阵：

A = np.array([[0.8, 0.2],
              [0.4, 0.6]])
B = np.array([0.4, 0.2])
C = np.array([0.28, 0.06])

我想做的是：

S = A.shape[0]
what_I_need = np.zeros([S])
for s in range(S):
    what_I_need[s] = np.sum(A[:, s] * C) * B[s]
print(what_I_need)

结果是

[0.0992 0.0184]

有没有for循环的优雅或pythonic方法吗？谢谢

Answer 1

使用matrix-multiplication-

A.T.dot(C)*B # or C.dot(A)*B

只需一次致电np.einsum-

np.einsum('ij,i,j->j',A,C,B) #use `optimize` flag for better perf.