Robust mixture regression based on the skew t distribution

TL;DR

This paper introduces a robust mixture regression method utilizing the skew t distribution to effectively model heavy-tailed and skewed errors, employing an EM algorithm for parameter estimation, validated through simulations and real data.

Contribution

It presents a novel mixture regression approach based on the skew t distribution with an EM algorithm for robust parameter estimation in skewed, heavy-tailed data.

Findings

The proposed method accurately models skewed and heavy-tailed errors.

Simulation studies demonstrate the estimator's robustness.

Real data application confirms practical effectiveness.

Abstract

In this study, we propose a robust mixture regression procedure based on the skew t distribution to model heavy-tailed and/or skewed errors in a mixture regression setting. Using the scale mixture representation of the skew t distribution, we give an Expectation Maximization (EM) algorithm to compute the maximum likelihood (ML) estimates for the paramaters of interest. The performance of proposed estimators is demonstrated by a simulation study and a real data example.