Monotone empirical Bayes test for uniform distributions using the maximum likelihood estimator of a decreasing density (Q578773): Difference between revisions

A monotone empirical Bayes test is derived for testing \(\theta \leq \theta_ 0\) against \(\theta >\theta_ 0\) given observations from a uniform distribution on [0,\(\theta)\). The test is based on the least concave majorant of the empirical distribution function. The test is shown to be asymptotically optimal, a Monte Carlo simulation exercise is presented to illustrate the validity of the asymptotic results for various values of the sample size.