Bayes and classical estimation of environmental factors for the binomial distribution

An environmental factor converts reliability test results at one environmental condition into equivalent "failure" information at other environments. This paper studies environmental-factor estimation for the binomial distribution. Under general conditions, Bayes point estimates and credibility limits for environmental factors are derived. Classical point and confidence interval estimates are introduced and compared with the Bayes estimators. The characteristics of Bayes and classical estimators for the binomial distribution are summarized through numerical computation and theoretical analysis. A numerical example of reliability assessment by means of environmental factors is presented.