Comparison between empirical mode decomposition and local linear quantile regression in the presence of boundaries

Abstract

Quantile regression has been explored extensively by many researchers in the last two decades. This branch of regression provides a thorough and satisfactory approach to the analysis of regression models. Nonparametric quantile regressions can sometimes produce much valuable information concerning departures from standard model assumptions (like nonlinearity and heteroscedasicity). However, a novel method known as empirical mode decomposition (EMD) is capable of providing a good description of the non-stationary and nonlinear relationships among random variables. This study compares nonparametric quantile regressions (local linear quantile regressions) and EMD in the presence of boundaries. A simulation study is conducted to assess the numerical result.