A bivariate Pareto model

被引:17
|
作者
Sankaran, P. G. [1 ]
Kundu, Debasis [2 ]
机构
[1] Cochin Univ Sci & Technol, Dept Stat, Cochin 682016, Kerala, India
[2] Indian Inst Technol, Dept Math & Stat, Kanpur 208016, Uttar Pradesh, India
关键词
Pareto distribution; bivariate hazard rate; copula; maximum-likelihood estimators; competing risks; COMPETING RISKS; DISTRIBUTIONS; IDENTIFIABILITY; RELIABILITY;
D O I
10.1080/02331888.2012.719521
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Lindley-Singpurwalla's [Multivariate distribution for the life lengths of a system sharing a common environment, J. Appl. Probab. 23 (1986), pp. 418-431] bivariate Pareto distribution is one of the most popular bivariate Pareto distributions. Sankaran and Nair [A bivariate Pareto model and its applications to reliability, Naval Res. Logist. 40 (1993), pp. 1013-1020] proposed a new bivariate Pareto distribution having Pareto marginals and containing Lindley-Singpurwalla's bivariate Pareto model as a special case. It also has several other interesting properties. In this paper, we re-visit Sankaran and Nair's bivariate Pareto model. We discuss several other new properties. The maximum-likelihood estimators and two-stage estimators are also investigated. We analyse two data sets for illustrative purposes. It is observed that this model can be used quite effectively for analysing competing risk data. Finally, we propose some generalizations.
引用
收藏
页码:241 / 255
页数:15
相关论文
共 50 条