Multivariate analysis

Hi my friends,

I was wondering how could I define a random variable in a matrix shape such that:
1 - Let A be such matrix mxn;
2 - Let the row vector B_1xn be another r.v. ;
3 - The data D has the same shape as A, and the likelihood is such that “B*A” is the mean of a multivariate binomial;

Each row of A is a different scenario of the same enviroment and I would like to proceed a joint analysis where each component of B (bi) behaves as a different scale parameter to each scenario (–A_i–).

Thank you guys!

I think I figured out. We can’t do a straight matrix multiplication as usual, you need to proceed a set_subtensor for each row of A and element of B. Thank you guys!

1 Like