Random coefficient regressions: parametric goodness-of-fit tests

被引：3

作者：

Delicado, PF

Romo, J

机构：

[1] Univ Carlos III Madrid, Dept Estadist & Econometria, Madrid, Spain

[2] Univ Politecn Catalunya, Dept Estadist & IO, E-08028 Barcelona, Spain

来源：

JOURNAL OF STATISTICAL PLANNING AND INFERENCE | 2004年 / 119卷 / 02期

关键词：

goodness-of-fit; linear regression; random coefficient; parametric empirical processes;

D O I：

10.1016/S0378-3758(02)00484-6

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

Random coefficient regression models have been applied in different fields during recent years and they are a unifying frame for many statistical models. Recently, Beran and Hall (Ann. Statist. 20 (1992) 1970) raised the question of the nonparametric study of the coefficients distribution. Nonparametric goodness-of-fit tests were considered in Delicado and Romo (Ann. Inst. Statist. Math. 51 (1999) 125). In this nonparametric framework, the study of parametric families for the coefficient distributions was started by Beran (Ann. Inst. Statist. Math. (1993) 639). Here we propose statistics for parametric goodness-of-fit tests and we obtain their asymptotic distributions. Moreover, we construct bootstrap approximations to these distributions, proving their validity. Finally, a simulation study illustrates our results. (C) 2002 Elsevier B.V. All rights reserved.

引用

页码：377 / 400

页数：24

共 26 条

[21]

Scheffe H., 1959, The analysis of variance

[22]

Shorack GR., 1986, Empirical processes with applications to statistics

[23] EFFICIENT INFERENCE IN A RANDOM COEFFICIENT REGRESSION MODEL [J].