w-hardy

AHP appears to ignore models with negative IC values

1

When some models have negative information criteria (e.g., BIC) values, the AHP appears to select the best model that has positive IC values, thereby ignoring better fitting models, see example...