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 scoring of date-based targets #102

Open
matthewcornell opened this Issue Oct 10, 2018 · 0 comments

Comments

1 participant
@matthewcornell
Member

matthewcornell commented Oct 10, 2018

Currently we only score 'step-ahead' targets such as '1 wk ahead'. More generally we can score any continuous ones (e.g., 'Season peak percentage'). However, we should also score non-continuous ones like 'Season onset' or 'Season peak week'. @nick explained to me that this can get complicated. We think the solution is to add support for a fixed set of Zoltar-standard time representations, and use existing libraries like pymmwr to implement them. We'd probably use Project.time_interval_type for this.

This came up b/c these date-based targets - Season onset and Season peak week - currently have truth values (which were generated by abhinav's script) that have been converted from EWs to timezero dates. However, the predicted values are still EWs.

@matthewcornell matthewcornell created this issue from a note in Iterations (Queue) Oct 10, 2018

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