Hello. In loess regression (or gam with cubic spline smoothers, I think) it is possible to fit models with different numbers of equivalent parameters – thus model df –and then conduct an inferential test via anova. Is this a valid way of choosing the smoother df?
Specifically, I fix a significance level of alpha and then fit a sequence of models with increasing numbers of model df (say 2,3,4…). I conduct an anova to compare this sequence of models and choose the smoother df as the one at which models fit with further increases do not result in a significant improvement. If this is not an acceptable strategy, what would people recommend beyond using the built in cross-validation criterion? Thanks for any leads. Bill Shipley Département de biologie, Université de Sherbrooke, Sherbrooke (Québec) J1K 2R1 CANADA [EMAIL PROTECTED] <http://callisto.si.usherb.ca:8080/bshipley/> http://callisto.si.usherb.ca:8080/bshipley/ [[alternative HTML version deleted]] ______________________________________________ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
