-
Notifications
You must be signed in to change notification settings - Fork 78
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.
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
LinearAlgebra.LAPACKException(3) ERROR when calling confidence_intervals
#181
Comments
version: |
I would guess singular or near-singular information matrix (or whatever equivalent is used here). |
When will I ever learn to read the provided example before answering. The problem is exactly that. Your third parameter gives you a zero in the hessian of the likelihood function (the interpretation we use when we think about standard errors). Your third gradient entry is exactly, constantly zero so your last row and column of your hessian is too.
something like that. In other words, the third parameter is unidentified. It could be anything. https://en.wikipedia.org/wiki/Identifiability |
Thank you. Although a bit too late for when I was using this, nevertheless I appreciate the answer. So to conclude, I shouldn't use "null" parameters and always use as many parameters as I actually use in my model. A suggestion for LsqFit: to throw a warning when this happens? If I read your message correctly, the package has all the information it needs to make such a statement? |
I knew it from your model. I guess it can be a bit hard to check. But if the inv (should be a factorization, but...) fails, we could certainly catch it and ask the user to check for identification, yes. |
Hi,
I'm running code that guaranteed worked just a week ago, but since then I've done some package updates, and also moved to Julia 1.6. Julia 1.6 is not the problem, same error I get when falling back to Julia 1.5.1.
Here is a MWE:
Any ideas how to fix?
p.s.: While making a MWE, I've checked the Tests folder of this package, and couldn't find a call to
confidence_interval
anywhere... Where is this function tested?The text was updated successfully, but these errors were encountered: