The model identification of beta distribution based on quantiles

Author: Shih Nenghui  

Publisher: Taylor & Francis Ltd

E-ISSN: 1563-5163|85|10|2022-2032

ISSN: 0094-9655

Source: Journal of Statistical Computation and Simulation, Vol.85, Iss.10, 2015-07, pp. : 2022-2032

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

This paper addresses the modelling of a beta distribution based on quantile pairs. The elicited beta distribution is proved to exist uniquely for any given two-quantile pairs. By combining the results of this study with the research documented in the literature, we concluded that for any given two-quantile pairs, there is one and only one beta distribution fits the two specified quantile pairs. The numerical results that verify this finding were obtained by comprehensively examining 1300 two-quantile pairs. The first four moments of the fitted beta distribution were also derived numerically using these two-quantile pairs. The variations in the first four moments, which reflect the perception of uncertainty, were plotted and were explored.