The problem of obtaining prediction intervals for the Rayleigh distribution is considered. Classical (frequentist) prediction intervals are obtained by construction of a pivotal quantity that has a beta distribution. The proposed intervals are compared to highest posterior density prediction intervals in the literature. The problem of minimum sample size determination is addressed for measuring the variability of the coverage of the β-expectation tolerance intervals. Tables for determination of sample size are provided
Published in:
Reliability, IEEE Transactions on
(Volume:40
,
Issue:
3
)
Date of Publication: Aug 1991