SELECTION OF VARIABLES BASED ON NONCONCAVE PENALIZED LIKELIHOOD USING LASSO, ELASTIC NET, AND SCAD METHOD
Femmy Diwidian (1*), Khairil A Notodiputro (2), Bagus Sartono (2)

(1) UIN Syarif Hidayatullah Jakarta
(2) IPB University
*Corresponding author: femmy.diwidian[at]uinjkt.ac.id

Abstract

Variable selection is an essential topic in linear regression analysis to improve predictability and to select significant variables. Estimating the regression coefficient on high-dimensional data cannot be done using the least squares method, so it requires specific analytical techniques. Approaches that can take on high-dimensional data include SCAD, LASSO, and Elastic Net. This research will analyze the most crucial method between SCAD, LASSO, and Elastic Net on Low Birth Weight (LBW) data in East Nusa Tenggara (NTT). Two methods are used in this study, first, comparing the SCAD, LASSO, and Elastic Net methods using simulation data, and the second applying the logistic regression method to actual data. The data used in this study is the LBW data by fertile women in NTT from the 2017 IDHS (Indonesian Demographic and Health Survey) data. The analysis shows that the results obtained through simulation and data reveal that SCAD is better than the other methods.

Keywords: Variable Selection, Nonconcave penalized likelihood, LASSO, Elastic net,

Topic: Mathematics