Some estimation techniques in reliability and survival analysis based on record-breaking data

Abstract

In this paper we review some of the classical and Bayesian results on statistical inference from records that can be used in reliability and survival analysis. We focus on some important lifetimemodels, giving special attention to heavy-tailed distributions in order to consider applications of record-breaking data to the study of extreme events. Results on the estimation of the number of observations needed to attain a given number of records are also studied in depth. Numerical illustrations and results on the estimation of cost functions are included as well. This chapter can serve as a guide for people interested in making inferences in the fields of reliability and survival analysis when only record values are available.