Handling chng "missingness"

cmu-delphi / exploration-tooling

tools for evaluating and exploring forecasters

Other

0 stars 0 forks source link

Open dsweber2 opened 6 months ago

dsweber2 commented 6 months ago

if we're trying to justify including change at all, having the "ideal" version (without contract induced data gaps) of the signal as a different covariate would be useful. This could be a different data target called e.g. chng_ideal. This means rewriting the history using historic lag patterns, but with "future" data (which is only latent because of logistics/contracts/etc).
imputation/extrapolation? maybe dangerous

dsweber2 commented 6 months ago

could execute the first of these is by imputation from the eventual as_of versions that are present, using historic revision patterns.