It is shown that a popular variable choice method of Hellwig, which is recommended in the Polish econometric textbooks does not enjoy a very basic consistency property. It means in particular that the method may lead to rejection of significant variables in econometric modeling. A simulation study and a real data analysis case are given to support theoretical results.
@article{bwmeta1.element.bwnjournal-article-doi-10_7151_dmps_1106,
author = {Tadeusz Bednarski and Filip Borowicz},
title = {On inconsistency of Hellwig's variable choice method in regression models},
journal = {Discussiones Mathematicae Probability and Statistics},
volume = {29},
year = {2009},
pages = {41-51},
zbl = {1208.62200},
language = {en},
url = {http://dml.mathdoc.fr/item/bwmeta1.element.bwnjournal-article-doi-10_7151_dmps_1106}
}
Tadeusz Bednarski; Filip Borowicz. On inconsistency of Hellwig's variable choice method in regression models. Discussiones Mathematicae Probability and Statistics, Tome 29 (2009) pp. 41-51. http://gdmltest.u-ga.fr/item/bwmeta1.element.bwnjournal-article-doi-10_7151_dmps_1106/
[000] [1] H. Akaike, Information theory and an extension of the maximum likelihood principle, In Second International Symposium on Information Theory, Budapest: Academiai Kiado (1973), 267-81. | Zbl 0283.62006
[001] [2] T. Bednarski and E. Mocarska, On robust model selection within the Cox model, Econometrics Journal 9, (2006), 279-290. | Zbl 1096.62138
[002] [3] Z. Hellwig, Problem optymalnego doboru predyktant, Przegląd Statystyczny nr 4 (1969).
[003] [4] J.A.F. Machado, Robust model selection and M-estimation, Econometric Theory 9 (1993), 478-493.
[004] [5] G. Schwarz, Estimating the dimension of a model, The Annals of Statistics 6 (2) (1978), 461-64.
[005] [6] D. Serwa, Metoda Hellwiga jako kryterium doboru zmiennych do modeli szeregów czasowych, Szkoła Główna Handlowa, Kolegium analiz Ekonomicznych, Instytut Ekonometrii 2004.