Generalized linear failure rate power series distribution

被引:7
|
作者
Harandi, S. Shams [1 ]
Alamatsaz, M. H. [1 ]
机构
[1] Univ Isfahan, Dept Stat, Esfahan, Iran
关键词
Compounding; EM algorithm; Failure rate function; Maximum entropy identification; Maximum likelihood estimation; mixing; LIFETIME DISTRIBUTION; WEIBULL EXTENSION;
D O I
10.1080/03610926.2013.878357
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
In this article, we introduce a new class of lifetime distributions. This new class includes several previously known distributions such as those of Chahkandi and Ganjali (2009), Mahmoudi and Jafari (2012), and Nadarajah etal. (2012). This new class of four-parameter distributions allows for flexible failure rate behavior. Indeed, the failure rate function here can be increasing, decreasing, bathtub-shaped or upside-down bathtub-shaped. Several distributional properties of the new class including moments, quantiles and order statistics are studied. An EM algorithm for computing the estimates of the parameters involved is proposed and some maximum entropy characterizations are discussed. Finally, to show the flexibility and potential of the new class of distributions, applications to two real data sets are provided.
引用
收藏
页码:2204 / 2227
页数:24
相关论文
共 50 条