Log-symmetric regression models for correlated errors with an application to mortality data

TL;DR

This paper introduces a new class of log-symmetric regression models that handle correlated errors, offering greater flexibility for modeling positive, asymmetric data, demonstrated through simulations and a mortality data application.

Contribution

The paper develops a novel log-symmetric regression model for correlated errors, including estimation methods, properties, and real-world application, advancing existing models.

Findings

Models effectively handle correlated positive data.

Simulation shows reliable parameter estimation.

Application to mortality data demonstrates practical utility.

Abstract

Log-symmetric regression models are particularly useful when the response variable is continuous, strictly positive and asymmetric. In this paper, we proposed a class of log-symmetric regression models in the context of correlated errors. The proposed models provide a novel alternative to the existing log-symmetric regression models due to its flexibility in accommodating correlation. We discuss some properties, parameter estimation by the conditional maximum likelihood method and goodness of fit of the proposed model. We also provide expressions for the observed Fisher information matrix. A Monte Carlo simulation study is presented to evaluate the performance of the conditional maximum likelihood estimators. Finally, a full analysis of a real-world mortality data set is presented to illustrate the proposed approach.