search

hubverse-org / hubEnsemblesManuscript

https://htmlpreview.github.io/?https://github.com/Infectious-Disease-Modeling-Hubs/hubEnsemblesManuscript/blob/master/analysis/paper/hubEnsembles_manuscript.html
Other
1 stars 2 forks source link

correct statement about PI coverage #61

Closed elray1 closed 3 months ago

elray1 commented 3 months ago

in section 5 we say "WIS and PI coverage evaluate how well a probabilistic forecast captures that same observed value, prioritizing probabilistic forecasts that are more certain"

But PI coverage does not "prioritize" or reward forecasts that are more certain; it does not measure any aspect of a distribution's sharpness.

lshandross commented 3 months ago

Addressed by pull request #63