Variable Selection for Linear Regression Model with General Variance

Authors

  • Nuchjaree Lenghoe
  • Jiraphan Suntornchost

Keywords:

Variable selection, Bias-reduction, Adjusted R2

Abstract

Lahiri and Suntornchost (2015) proposed adjustments to variable selection criteria for the Fay-Herriot
model by considering sampling errors. Their results were shown to improve traditional variable selection criteria.
In this study, we extend their method to construct variable selection criteria for linear regression models
with general variance assumption allowing for the possibility of correlated regression errors. Closed forms of
statistics for variable selection criteria are provided with numerical studies. Simulation results show that our
proposed variable selection criteria can reduce the approximation errors of the standard variable selection criterion.

Author Biographies

Nuchjaree Lenghoe

Mathematics and Computer Science, Chulalongkorn University, Bangkok

Jiraphan Suntornchost

Mathematics and Computer Science, Chulalongkorn University, Bangkok

References

[1] Burnham KP, Anderson DR. Model selection and multimodel inference: a practical informationtheoretic
approach. 2nd ed. New York: Springer; 2002.

[2] Casella G, Berger RL. Statistical inference. 2nd ed. Pacific Grove, CA: Duxbury; 2002.

[3] Chatterjee S, Lahiri P. A simple computational method for estimating mean squared prediction error
in general small-area model: Proceedings of the section on survey research methods; American Statistical
Association Alexandria VA; 2007:3486-3493.

[4] Hurvich CM, Tsai CL. Regression and time series model selection in small samples. Biometrika.
1989:297-307.

[5] Lahiri P, Suntornchost J. Variable selection for linear mixed models with applications in small area
estimation. Sankhya B. 2015; 77(2):312-20.

[6] Magnus JR, Neudecker H. Matrix differential calculus with applications in statistics and econometrics.
3nd Netherlands: John Wiley & Sons; 2007.

[7] Neath AA, Cavanaugh JE. The Bayesian information criterion: background, derivation, and applications.
Wiley Interdisciplinary Reviews: Computational Statistics. 2012; 4(2):199-203.

[8] Rencher AC, Schaalje GB. Linear models in statistics. 2nd ed. Canada: John Wiley & Sons; 2008.

Downloads

Published

2018-06-30

How to Cite

Lenghoe, N., & Suntornchost, J. (2018). Variable Selection for Linear Regression Model with General Variance. Journal of Applied Statistics and Information Technology, 2(2), 15–24. Retrieved from https://ph02.tci-thaijo.org/index.php/asit-journal/article/view/165718