Basu-Dhar's bivariate geometric distribution in presence of censored data and covariates: some computational aspects

Some computational aspects to obtain classical and Bayesian inferences for the Basu and Dhar (1995) bivariate geometric distribution in presence of censored data and covariates are discussed in this paper. The posterior summaries of interest are obtained using standard existing MCMC (Markov Chain Monte Carlo) simulation methods available in popular free softwares as the OpenBugs software and the R software. Numerical illustrations are introduced considering simulated and real datasets showing that the use of discrete bivariate distributions may be a good alternative to the use of continuous bivariate distributions, in many areas of application.

本文探讨了在存在截尾数据与协变量的情境下，针对Basu与Dhar（1995）提出的二元几何分布开展经典推断与贝叶斯推断所需的若干计算层面问题。本文借助现有主流免费统计软件（如OpenBugs与R软件）中的标准马尔可夫链蒙特卡洛（Markov Chain Monte Carlo, MCMC）模拟方法，获取了目标后验汇总统计量。本文通过模拟数据集与真实数据集开展数值算例分析，结果表明在诸多应用领域中，离散型二元分布可作为连续型二元分布的优质替代方案。