Predicted Squared Error: A Criterion for Automatic Model Selection
Andrew R. Barron · 2020
Whereas conventional empirical modeling techniques require an assumed model structure, new procedures have been developed which generate the model structure as well as the model coefficients from a data base. These procedures include the GMDH and PNETTR algorithms for creating polynomial networks. Key to any automatic procedure for generating models is the criterion for ranking different model structures and selecting the best.