An algebraic aspect of Pareto mixture parameter estimation using censored sample: A Bayesian approach

J Integr Neurosci. 2018;17(3-4):463-477. doi: 10.3233/JIN-180082.

Abstract

Applications of Pareto distribution are common in reliability, survival and financial studies. In this paper, A Pareto mixture distribution is considered to model a heterogeneous population comprising of two subgroups. Each of two subgroups is characterized by the same functional form with unknown distinct shape and scale parameters. Bayes estimators have been derived using flat and conjugate priors using squared error loss function. Standard errors have also been derived for the Bayes estimators. An interesting feature of this study is the preparation of components of Fisher Information matrix.

Keywords: Bayes estimators; Fisher information matrix; Maximum likelihood estimators; informative priors; predictive intervals; probabilistic mixing; right censoring; uninformative priors.