Summary
A general frame is given in which model selection criteria can be derived. At first conditions are stated under which the asymptotic distribution of minimum discrepancy estimators can be given. The results are used to derive an approximation to the expected discrepancy. Estimators of this expectation, i.e. model selection criteria, are proposed.
Zusammenfassung
Ein allgemeiner Rahmen wird gegeben, in dem man Kriterien zur Modellauswahl herleiten kann. Zuerst werden Bedingungen angeführt, unter denen man die asymptotische Verteilung von Minimum-Diskrepanz-Schätzern angeben kann. Mit den Resultaten wird dann eine Annäherung an die erwartete Gesamtdiskrepanz berechnet. Schätzer dieser Erwartung, also Kriterien zur Modellauswahl, werden vorgeschlagen.
Similar content being viewed by others
References
Akaike H (1969) Fitting autoregressive models for prediction. Ann Inst Statist Math 21:243–247
Akaike H (1973) Information theory and an extension of the maximum likelihood principle. In: Petrov BN, Czaki F (eds) Proceedings of the 2nd International Symposium on Information Theory, Akademiai Kiado, Budapest, pp 267–281
Jennrich RI (1969) Asymptotic properties of non-linear least squares estimators. Ann Math Statist 40:633–643
Linhart H, Zucchini W (1982) On model selection in analysis of variance. In: Fleischmann B et al. (eds) Operations Research Proceedings 1981. Springer, Berlin Heidelberg New York, pp 483–493
Linhart H, Zucchini W (1982) A method for selecting the covariates in analysis of covariance. S Afr Statist J 16:97–112
Parzen E (1974) Some recent advances in time series modelling. Trans Auto Control AC-19:723–730
White H (1982) Maximum likelihood estimation of misspecified models. Econometrica 50:1–25
Zwanzig S (1980) The choice of approximative models in nonlinear regression. Math Operationsforschung Statist Ser Statistics 11:23–47
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Linhart, H., Volkers, P. Asymptotic criteria for model selection. OR Spektrum 6, 161–165 (1984). https://doi.org/10.1007/BF01719613
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/BF01719613