scientific article; zbMATH DE number 3874459

From MaRDI portal
Publication:3339993





[For the entire collection see Zbl 0521.00020.] This chapter contains a summary of the statistical theories of model selection. Sections 2 and 3 include most of the traditional model selection problems. Section 2 deals with alternative models which spring from a priori judgment. It is maintained that a constrained model might be either true or approximately true. Classical hypotheses testing is discussed, as well as Bayesian treatments with complete and incomplete prior distributions. Subsections deal also with measuring the multicollinearity problem and with inference given zero degrees of freedom. Models in Section 3 arise not from the judgment of the investigator as in Section 2, but from his purpose, which is measured by a formal loss function. Quadratic loss functions are considered, both with and without fixed costs. A brief comment is made about ridge regression. The main conclusion of Section 3 is that quadratic loss does not imply a model selection problem. Sections 4, 5, and 6 discuss problems which are not as well known. The problem of selecting the best proxy variable is treated in Section 4. Akaike's information criterion is discussed in Section 5, although it is pointed out that, except for subtle conceptual differences, his problem reduces to estimation with quadratic loss. Section 6 deals with methods for discounting evidence when models are discovered after having previewed the data. Finally, Section 7 contains material on stepwise regression, cross-validation, and goodness-of-fit tests.




Cited in
(33)








This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3339993)