Dynamic models for statistical inference from accelerated life tests

Abstract

An approach is presented for inference from accelerated life tests. The approach is based on a dynamic linear model which arises naturally from the accelerated life testing problem and uses linear Bayesian methods for inference. The advantage of the procedure is that it does not require large numbers of items to be tested and that it can deal with both censored and uncensored data. Furthermore, the approach produces closed-form inference results. The use of the approach with some actual accelerated life test data is illustrated.<>