A unified treatment of multiple testing with prior knowledge using the p-filter (Q2328060)

The problems related with the multiple hypothesis testing are considered. Given a large set of null hypothesis, the goal of multiple testing is to decide which subsets should be rejected, while guaranteeing control on the number of false rejections. The authors consider methods that control the \textit{false discovery rate} or test the \textit{global null} hypothesis while incorporating a number of the following strategies related with prior knowledge: the use of prior weights, the use of penalty weights, the partition hypotheses into groups, the dependence structure within the data. The authors of the paper focus on the conceptual and mathematical aspects of multiple decision-making with prior knowledge.