product of 2 csr matrices with correct shapes

Question

I am required to multiply 2 csr matrices with shapes A : (385019, 72) B : (72, 385019). I do see that the # of columns of A is 72 and thats exactly the number of rows in B. Yet when I perform sparse.csr_matrix(A).multiply(sparse.csr_matrix(A)) I get ValueError: inconsistent shapes I have been through other posts but nothing has helped me yet. Very large matrices using Python and NumPy

score 6 · Accepted Answer · answered Feb 14 '18 at 16:40

6

The multiplication you are looking for is called "dot product" and in python you can do that as follows

sparse.csr_matrix(A) * sparse.csr_matrix(B)

However, the multiplication that you are using sparse.csr_matrix(A).multiply(sparse.csr_matrix(A)) in the problem you described is called "Point-wise multiplication by another matrix, vector, or scalar". This means that every element of A will be multiplied by every element of B if both A and B are matrices; in this case size of A and B must be same. If B is a scalar then every element of A will be multiplied by B.

answered Feb 14 '18 at 16:40

Mohsin

536
6
18

2

Thanks a lot Mohsin :) That helped. – vku Feb 14 '18 at 17:06
1

it was my pleasure to help – Mohsin Feb 14 '18 at 20:29

product of 2 csr matrices with correct shapes

1 Answers1