ISSN 2630-0583 (Print)

ISSN 2630-0656 (Online)

JCST

Journal of Current Science and Technology

http://jcst.rsu.ac.th

Rangsit Journal of Arts and Sciences. Vol.3 No.2 , July - December 2013.

A variable selection method in multiple linear regression models based on Tabu Search

Kannat Na Bangchang

Abstract

         This research has proposed a variable selection method based on the Tabu Search for multiple linear regression models.  In this study two objective functions used in the Tabu Search are mean square error (MSE) and the mean absolute error (MAE).  The results of the Tabu Search are compared with the results obtained by the stepwise regression method based on the hit percentage criterion.  The simulations cover both cases, without and with multicollinearity problems.  Without multicollinearity problem, the hit percentages of the stepwise regression method and the Tabu Search using the objective function of MSE are almost the same but slightly higher than the Tabu Search using the objective function of the MAE.  But with multicollinearity problem the hit percentages of the Tabu Search using the objective function of MSE or the MAE are higher than the hit percentage of the stepwise regression method. Additionally, the correlation coefficients between the independent variables X1 and X4 are higher; yielding hit percentages that are lower.

Keywords: stepwise regression, Tabu Search, variable selection

Download Full Paper.