On convergence of indicator-based estimators for parameters of linear model
The problem of parameters estimation of linear model is considering for the case, when random errors are independent andtheir common distribution has known joint quintiles of a given levels. The uniform law of large numbers and consistency has beenproved for a class of indicator-based estimators. The proposed estimator is applicable for utilizing prior information and particularlyto the case of quantile regression.
Download file
Counter downloads: 253
Keywords
Authors
| Name | Organization | |
| Tarassenko P.F. | Tomsk State University | ptara@ich.tsu.tomsk.su |
References
Болдин М.В., Симонова Г.И., Тюрин Ю.Н. Знаковый статистический анализ линейных моделей. М.: Наука, 1997.
Tarassenko P.F. Hypothesis testing for indicator analysis of linear models // KORUS-2002, 6th Russian-Korean International Symposium on Science and Technology, June 24-30, 2002, Novosibirsk: Proceedings. 2002. V.3, Mathematics. P.185.
Tarassenko P.F. On indicator-based hypothesis testing // 24th European Meeting of Statisticians (EMS2002), 14th Prague Conference on Information Theory, Statistical Decision Functions and Random Processes - Prague, August 19-23, 2002.
Тарасенко П.Ф. Оптимальные тесты, основанные на индикаторах событий // Вестник ТГУ. 2002. Прил. №1(I), сент. Докл. IV Всеросс. конф. «Нов. информ. технологии в иссл. сложн. структур». Томск, 10 - 13 сентября 2002. С.185-190.
Koenker R., Bassett G. Regression quantiles // Econometrica. 1978. V.46. P.33-50.
Koenker R., Portnoy S. M Estimation of multivariate regressions // JASA. 1990. V.85, issue 412. P.1060-1068.
Koenker R., Ng, P., Portnoy S. Quantile smoothing splines // Biometrika. 1994. V.81. No. 4. P.673-680.
