报告主题：Hierarchical Bayesian Inference for Tweedie Exponential Dispersion Process with Random Drifts

时间：2024年5月31日15:00-16:00

地点：上川路校区一教218室

报告人：王平平

报告内容简介：

The stochastic process models have become the most popular way to model degradation data for high-quality products. However, if the model is mis-specified, we shall obtain poor reliability assessment. Since Tweedie exponential dispersion (TED) process contains three classical stochastic processes as special cases, the application of TED process transforms the model selection problem into a parameter estimation problem dexterously. In this paper, we propose a TED process with random drifts to model degradation data under usage stress and a TED process with random drifts and covariates to model accelerated degradation data, where random drifts are considered to explain heterogeneous degradation rates. A hierarchical Bayesian method is adopted to estimate the parameters in our proposed models by using Gibbs algorithm. We also derive the failure-time distribution and the remaining useful life (RUL) distribution for our proposed models. A variety of simulation studies show that our proposed models outperform the wrongly specified models and are very close to the right model. Two illustrative examples demonstrate that the proposed TED process with random drifts and the TED process with random drifts and covariates can give rise to more suitable fits to the degradation data and accelerated degradation data.

主讲人简介：

王平平，南京财经大学讲师，博士毕业于华东师范大学，获统计学理学博士学位，发表SCI等核心期刊十余篇，主持国家自然科学基金一项，江苏省自然科学基金一项，教育部人文社会科学基金一项。研究方向为可靠性退化数据分析与空间统计分析方法。