This is my first uploading in this forum. I’m now trying to implement simple linear regression task, but it is slightly different from common ols problem.
I need a help to implement first term efficiently. Is there any module that help computing weighted least square?
My trial is
output = W.matmul(X) # k * D matrix l = torch.diagonal((output-y).T.matmul(Sigma).matmul(output-y)).sum()
Thank you in advance,