B0492
Title: Adaptive L0 approach for sparse quantile regression
Authors: Christou Antonis - Cardiff University (United Kingdom)
Andreas Artemiou - University of Limassol (Cyprus) [presenting]
Abstract: The use of an adaptive L0 penalty is proposed in the quantile regression setting. It is demonstrated that the sparse estimator is found using a quadratic optimization procedure which demonstrates the equivalence between quantile regression and support vector regression. The performance of the method is compared to LASSO and adaptive LASSO in simulated and real settings and demonstrates the competitiveness of the new approach.
