系统工程与电子技术

• 可靠性 • 上一篇    下一篇

基于竞争失效数据的Lindley分布参数估计

黄文平1, 周经伦1, 宁菊红2, 金光1   

  1. 1. 国防科学技术大学信息系统与管理学院, 湖南 长沙 410073;
    2. 江西师范大学数学与信息科学学院, 江西 南昌 330027
  • 出版日期:2016-01-30 发布日期:2010-01-03

Parameter estimation of Lindley distribution with competing risk data

HUANG Wen-ping1, ZHOU Jing-lun1, NING Ju-hong2, JIN Guang1   

  1. 1. College of Information System and Management, National University of Defense Technology, Changsha 410073, China; 2. College of Mathematics and Information Science, Jiangxi Normal University, Nanchang 330027, China
  • Online:2016-01-30 Published:2010-01-03

摘要:

假定产品在使用时经历多种竞争失效,每种失效寿命数据服从Lindley分布。竞争失效数据包括不完全数据和截尾数据。对竞争失效分布进行参数点估计、相对风险率计算,讨论了参数的渐进置信区间和Bootstrap置信区间。为考察其适应性,一组实际数据用来说明各种参数估计方法的使用情况,同时,分别用指数分布和威布尔分布拟合该组数据,并计算相应的统计量。结果表明,根据3种分布的极大似然估计量和K-S值,Lindley分布对该组数据具有最好的适应性。

Abstract:

The products are supposed to fail with competing risks, and each risk conforms to the Lindley distribution.The competing risks fail data include incomplete data and censoring data.The maximum likelihood procedure is used to derive point estimation of the unknown parameters, and their asymptotic confidence interval estimation and Bootstrap confidence interval estimation are discussed.The relative risks due to each cause of failure are investigated. The theoretical results obtained on a set of real data are applied.Also, hypothesis tests are studied to investigate whether the real data set can be fitted well by the Lindley distribution. The results show that the Lindley distribution fits best compared with the exponential distribution and Weibull distribution.