An SVM-like Approach for Expectile Regression

TL;DR

This paper introduces a support vector machine-like method for expectile regression, providing an efficient solver and demonstrating its effectiveness through experiments and comparisons with existing tools.

Contribution

It develops a novel SVM-like approach for expectile regression and proposes a sequential minimal optimization-based solver for the optimization problem.

Findings

The solver is efficient and scalable.

Experimental results show competitive performance.

Comparison with ER-Boost demonstrates advantages in certain scenarios.

Abstract

Expectile regression is a nice tool for investigating conditional distributions beyond the conditional mean. It is well-known that expectiles can be described with the help of the asymmetric least square loss function, and this link makes it possible to estimate expectiles in a non-parametric framework by a support vector machine like approach. In this work we develop an efficient sequential-minimal-optimization-based solver for the underlying optimization problem. The behavior of the solver is investigated by conducting various experiments and the results are compared with the recent R-package ER-Boost.