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

Add softmax to broadcast_dayofweek #258

Closed
seabbs opened this issue Jun 4, 2024 · 2 comments · Fixed by #260
Closed

Add softmax to broadcast_dayofweek #258

seabbs opened this issue Jun 4, 2024 · 2 comments · Fixed by #260

Comments

@seabbs
Copy link
Collaborator

seabbs commented Jun 4, 2024

#253 added a new helper to construct day of the week ascertainment models.

In the review of that PR (and elsewhere) it was noted that it might make sense to constrain the within week model to sum to 7 using a LogExpFunctions.softmax. This makes sense to provide as the default I think.

To do this we would need to add a TransformLatentModel wrapper to the latent_model passed to broadcast_dayofweek.

@SamuelBrand1
Copy link
Collaborator

This is a nice functionality, but should it be default? As mentioned in #253 the current implementation (some daily scale application) is probably the most common form of weekly periodic effect in observations used in modelling.

@seabbs
Copy link
Collaborator Author

seabbs commented Jun 4, 2024

I think it is effectively the same but constrained (at least how I am thinking about it) so it seems reasonable to make it a default in this kind of helper (in the world of implementing helpers for good ideas).

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

Successfully merging a pull request may close this issue.

2 participants