ESTIMATING AND SELECTING VARIABLES FOR A CONDITIONAL QUANTILE REGRESSION MODEL USING PENALTY FUNCTIONS

Mohammed Ibrahim Zamel (Dhi-Qar Education Directorate, Iraq)

Jurnal Ekonomi, Bisnis dan Pendidikan (JEBP)
Vol. 6 No. 6 (2026)

Penalised quantile regression is an optimal method for variable selection when confronted with a substantial number of explanatory variables. Certain factors may exert no influence on the response variable, thereby complicating the model and rendering it challenging to interpret. This research concentrates on variable selection utilising punitive functions (lasso) and (MCP), and evaluates them according to the mean square error criterion (MSE) for estimate purposes. Utilising the metrics of false positive rate (FPR) and false negative rate (FNR) to evaluate the estimator's efficacy in variable selection, simulated experiments revealed that the MCP estimator excels in both estimation and variable selection.