in case of fitting to multiple types of observational data, need to make sure date indices for observed and forecast dates are data-specific - so the index making bits probably should go inside the observational data likelihood definition wrappers, and be directly made from named date sequences, not from position indices in a date sequence
AugustHao
commented
11 months ago
in case of fitting to multiple types of observational data, need to make sure date indices for observed and forecast dates are data-specific - so the index making bits probably should go inside the observational data likelihood definition wrappers, and be directly made from named date sequences, not from position indices in a date sequence