Closed aksarkar closed 2 years ago
https://github.com/stephenslab/susieR/blob/a9883dabe94e499f2a5c0e7d4988db61766048dc/R/susie_utils.R#L670
I think this should be
0.5 * sum((1-s)*eigenld$values + s) +
since the determinant is the product of the eigenvalues, and this term involves the log determinant.
I am wrong about this. But the objective function still looks badly behaved when evaluated on a grid, so something is wrong.
https://github.com/stephenslab/susieR/blob/a9883dabe94e499f2a5c0e7d4988db61766048dc/R/susie_utils.R#L670
I think this should be
since the determinant is the product of the eigenvalues, and this term involves the log determinant.