Avoid cholesky decomposition of diagonal matrices in linear regression models #488

jseabold opened this Issue Oct 1, 2012 · 0 comments


None yet

1 participant

jseabold commented Oct 1, 2012

We needlessly take the cholesky decomposition of pinv(sigma) even in cases where we know sigma is diagonal. This could be avoided.

@jseabold jseabold closed this in d10372d Oct 5, 2012
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment