华东师范大学学报(自然科学版) ›› 2010, Vol. 2010 ›› Issue (5): 38-48.

• 应用数学与基础数学 • 上一篇    下一篇

退化数据分析的EM算法

徐安察, 汤银才   

  1. 华东师范大学 金融与统计学院, 上海 200241
  • 收稿日期:2010-01-01 修回日期:2010-05-01 出版日期:2010-09-25 发布日期:2010-09-25
  • 通讯作者: 汤银才

EM algorithm for degradation data analysis

XU An-cha, TANG Yin-cai   

  1. School of Finance and Statistics, East China Normal University, Shanghai 200241, China
  • Received:2010-01-01 Revised:2010-05-01 Online:2010-09-25 Published:2010-09-25
  • Contact: TANG Yin-cai

PDF

1511

摘要： 提出了用EM算法对产品可靠性进行分析. 在退化模型中, 当随机效应服从指数族分布时, 推导了参数估计的一般公式并且通过模拟评价了随机效应分布选取的敏感性. 最后分析了两个实际例子.

关键词: 退化数据, 随机效应, EM算法, 退化数据, 随机效应, EM算法

Abstract: Statistical inference procedures were proposed in assessing product reliability based on EM algorithm. The general formulae to estimate parameters were derived in the degradation model when its random effects follow an exponential family, and sensitivity of the choice of the distributions of the random effects was assessed via simulation. Two real data sets were analyzed.

Key words: random effect, EM algorithm, degradation data, random effect, EM algorithm

中图分类号: