AIC / BIC reported incorrectly when constantcoef used(?)


I am not sure where to report this but it seems that when specifying constantcoef's to a model fit, numparam in the aic and bic methods ignore the number of applied restrictions. The dimension of the transformed parameters is used, but this contains the constrained parameters as well.

Perhaps it is better to use the length of fpm.optim.minimizer?


Thanks for reporting this. It was indeed a bug which will be fixed in the 1.0 release.