Statistical inference for the semi-parametric proportional reversed hazard model for left-censored and zero-inflated data
-
Magdalena Pereda Vivo [1] ; Christian Paroissin [1]
-
[1] University of Pau and Pays de l'Adour
University of Pau and Pays de l'Adour
Arrondissement de Pau, Francia
-
- Localización: Test: An Official Journal of the Spanish Society of Statistics and Operations Research, ISSN-e 1863-8260, ISSN 1133-0686, Vol. 35, Nº. 1, 2026, págs. 157-179
- Idioma: inglés
- Texto completo no disponible (Saber más ...)
-
Resumen
When analysing concentrations obtained from analytical instruments, the resulting measurements are often subject to a limit of detection (LOD). In such cases, values reported below the LOD may correspond to either the true absence of the substance or positive concentrations that are left-censored. This work therefore aims to analyse data subject to left-censoring and zero inflation, where it is impossible to distinguish between true zeros and positive values that have been left-censored. Although some articles in the literature propose a mixture model for this type of data, they usually assume a parametric distribution for strictly positive values. In contrast, we propose a semi-parametric mixture model that assumes a proportional reversed hazard regression model for the positive part. We develop an expectation–maximization algorithm for parameter estimation and establish the asymptotic properties of the estimators. The methodology is validated through simulation studies and illustrated with an application to a real dataset.
