revise KDE fits and training set predictions to be prospective

reichlab / 2017-2018-cdc-flu-contest

Code and submissions for 2017-2018 CDC flu prediction contest

3 stars 2 forks source link

Closed nickreich closed 7 years ago

nickreich commented 7 years ago

[x] revise fit_region_kdes() function to only be doing prospective fitting (not LOSO).
[x] revise get_log_scores_via_direct_simulation() to be reading in prospective fits (around line 571 in kde-utils.R)
[ ] investigate whether to use trunc-KDE fits or not (which would involve loading real-time, unrevised data)

Note that we do not need to incorporate unrevised data into this model because it does not use real-time data (unless we use the trunc-kde approach).

nickreich commented 7 years ago

nickreich commented 7 years ago

Commit 9cfe06bc49d7b2bc951e7244ba2fecd5cbef848d fixes top two on list above.

nickreich commented 7 years ago

moving trunc-kde to its own issue.