Interval estimation of location and scale parameters based on record values (Q1382194)

The paper describes a method for constructing exact confidence intervals for the location, scale, and quantiles of a location-scale family of distributions based on upper record values. Prediction intervals for the future record based on the current record values are also considered. The intervals are derived by first obtaining the distributions of appropriate pivotal quantities conditioned on a vector of ancillary statistics. The method is illustrated by applying it to the extreme-value distribution. A numerical example is also given.