New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

audit Hogg hacks to avoid negative psi #13

Open

davidwhogg opened this issue Oct 17, 2012 · 0 comments

Assignees

Collaborator

davidwhogg commented Oct 17, 2012

There were three things happening to bring us negative psi values:

If we update mean before we update lambda and lambda before we update psi, we lose provable matrix thingies. If you know what I'm sayin.
When lambdas are HUGE and psis are tiny, numerical issues screw us.
When sum(self.rs[k]) gets small, life really SUCKS. I put in a floor on that. Yes, that is a horrible hack.

Search code for "HACK" and "MAGIC".

ghost assigned rossfadely

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment