Bayesian Nonparametric Statistical Inference for Shock Models and Wear Processes.
CALIFORNIA UNIV BERKELEY OPERATIONS RESEARCH CENTER
Pagination or Media Count:
Statistical procedures for shock models and wear processes are considered in this paper. We show that independent gamma-Dirichlet priors are conjugate priors when sampling from these shock models. Bayes rules given the observations are computed. In particular, we calculate the Bayes estimates of the survival probabilities for these models. We show consistency of the posterior distribution as well as weak convergence of the centered and suitably rescaled posterior processes. Author
- Statistics and Probability