Monotone false discovery rate

TL;DR

This paper introduces a method to produce monotone estimates of local and tail false discovery rates in large-scale testing, ensuring asymptotic optimality and good finite-sample properties.

Contribution

It presents a novel monotonization procedure for false discovery rates that is asymptotically optimal and effective in finite samples.

Findings

Monotonization improves false discovery rate estimates.

Method achieves asymptotic optimality.

Finite-sample properties are attractive.

Abstract

This paper proposes a procedure to obtain monotone estimates of both the local and the tail false discovery rates that arise in large-scale multiple testing. The proposed monotonization is asymptotically optimal for controlling the false discovery rate and also has many attractive finite-sample properties.