Mixture Representations and Inference for the Voigt Profile
Statistical Analysis and Data Mining: An ASA Data Science Journal · 2026-12 · 19 (6)
DOI: 10.1002/sam.70127 ↗Kurz gefasst
Fachartikel in der Zeitschrift Statistical Analysis and Data Mining: An ASA Data Science Journal, online erschienen am 17.11.2026. Verfasst von Massimo Cannas und 1 weiteren Autor:innen. Die Arbeit nennt 26 Literaturangaben.
Abstract (Originalsprache)
The Voigt profile is the density resulting from the convolution of a Gaussian and a Cauchy density. A representation of the Voigt profile in terms of elementary functions does not exist and the usual estimation approach is based on approximating the complex error function (Faddeeva) in the density before fitting the empirical profile. In this paper we discuss mixture characterizations of the Voigt profile which allow parameter estimation in a Bayesian framework without the need to approximate the Faddeeva function. Monte Carlo simulations show that the proposed approach outperforms traditional numerical fitting methods, achieving better Bayesian coverage, lower mean square error across all parameter combinations, and generally faster computation. The approach can be applied to empirical Voigt profile data and it is illustrated through a case study involving the dating of ancient artifacts based on Voigt spectral line‐shapes, which capture the chemical signatures of dyes present on their surfaces.
Zitieren
APA
Cannas, M., & Piras, N. (2026). Mixture Representations and Inference for the Voigt Profile. Statistical Analysis and Data Mining: An ASA Data Science Journal, 19(6). https://doi.org/10.1002/sam.70127
Vancouver
Cannas M, Piras N. Mixture Representations and Inference for the Voigt Profile. Statistical Analysis and Data Mining: An ASA Data Science Journal. 2026;19(6). doi:10.1002/sam.70127
BibTeX
@article{Cannas2026,
author = {Cannas, Massimo and Piras, Nicola},
title = {Mixture Representations and Inference for the Voigt Profile},
journal = {Statistical Analysis and Data Mining: An ASA Data Science Journal},
year = {2026},
volume = {19},
number = {6},
doi = {10.1002/sam.70127}
}
EUVIMED