Skip to content
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

Numerical precision problems #16

Open
DavideTisi opened this issue May 6, 2020 · 1 comment
Open

Numerical precision problems #16

DavideTisi opened this issue May 6, 2020 · 1 comment

Comments

@DavideTisi
Copy link

DavideTisi commented May 6, 2020

Hi, I want to spend two words on a possible source of problems in the multivariate cepstral analysis.
Using the multivariate methods, if you (wrongly) decorrelate from two linearly dependent currents, in some cases, it is possible that the code does not issue an error, and returns a result that seems reasonable, even though it is, of course, wrong.
The reason is that if the two currents are linearly dependent theoretically, but not exactly numerically, the resulting determinant of the covariate matrix is small but not exactly zero.

@lorisercole
Copy link
Member

Hi @DavideTisi!
This is an old issue, but it may still be relevant in SporTran 1.0 since the core modules have not changed.
I understand what happens, but it is quite hard to reproduce it. If you happen to have some data showing this problem, it would be very useful to draft a solution, or at least a check to be performed in such cases. Thanks

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

3 participants