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

Centers in dbaclust #67

Closed
jaksle opened this issue Aug 12, 2024 · 4 comments
Closed

Centers in dbaclust #67

jaksle opened this issue Aug 12, 2024 · 4 comments

Comments

@jaksle
Copy link
Contributor

jaksle commented Aug 12, 2024

I am trying to perform clustering based on DTW and I wanted to check dbaclust method here, I also need cluster (bary)centers. But the ones returned by dbaclust looked wrong. So I used method dba on the obtained clusters with the same DTW(250) as dbaclust, the result was different and much more sensible. I am at loss here. What are those centers if not dba barycenters?

@baggepinnen
Copy link
Owner

Have you seen the warning in the readme, and the links therein?
https://github.com/baggepinnen/DynamicAxisWarping.jl?tab=readme-ov-file#clustering-and-barycenter-averaging

DBA is known not to be a great algorithm at producing barycenters. See the example under https://github.com/baggepinnen/DynamicAxisWarping.jl?tab=readme-ov-file#differentiable-soft-dtw for how to produce much nicer barycenters

@jaksle
Copy link
Contributor Author

jaksle commented Aug 13, 2024

But this is not my question. My question is if this even is dba.

@baggepinnen
Copy link
Owner

The dba iteration code is called from within dbaclust, but with a smaller default number of iterations than the default of calling dba directly

@jaksle
Copy link
Contributor Author

jaksle commented Aug 13, 2024

Oh, this must be the source of the difference. Thank you!

@jaksle jaksle closed this as completed Aug 13, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants