Semiparametric two-component mixture models under L-moments constraints

TL;DR

This paper introduces a semiparametric two-component mixture model using L-moments constraints for the unknown component, offering better tail information handling and improved estimation via $\

Contribution

It proposes a novel mixture model incorporating L-moments constraints and develops feasible estimation algorithms with proven asymptotic properties.

Findings

L-moments constraints improve tail distribution estimation.

The proposed method outperforms moment-based approaches in simulations.

Estimation algorithms are validated through asymptotic analysis.

Abstract

We propose a structure of a semiparametric two-component mixture model when one component is parametric and the other is defined through L-moments conditions. Estimation of a two-component mixture model with an unknown component is very difficult when no particular assumption is made on the structure of the unknown component. A previous work was proposed to incorporate a prior linear information concerning the distribution function of the unknown component such as moment constraints. We propose here to incorporate a prior linear information about the quantile function of the unknown component instead. This information is translated by L-moments constraints. L-moments hold better information about the tail of the distribution and are considered as good alternatives for moments especially for heavy tailed distributions since they can be defined as soon as the distribution has finite expectation. The new semiparametric mixture model is estimated using $\varphi-$divergences which permit to build feasible algorithms. Asymptotic properties of the resulting estimators are studied and proved under standard assumptions. Simulations on data generated by several mixtures models demonstrate the viability and the interest of our novel approach and the gain from using L-moment constraints in comparison to the use of moments constraints.