False Discovery Rate Control under Archimedean Copula

We are considered with the false discovery rate (FDR) of the linear step-up test considered by Benjamini and Hochberg (1995). It is well known that controls the FDR at level if the joint distribution of -values is multivariate totally positive of order 2. In this, denotes the total number of hypotheses, the number of true null hypotheses, and the nominal FDR level. Under the assumption of an Archimedean -value copula with completely monotone generator, we derive a sharper upper bound for the FDR of as well as a non-trivial lower bound. Application of the sharper upper bound to parametric subclasses of Archimedean -value copulae allows us to increase the power of by pre-estimating the copula parameter and adjusting . Based on the lower bound, a sufficient condition is obtained under which the FDR of is exactly equal to , as in the case of stochastically independent -values. Finally, we deal with high-dimensional multiple test problems with exchangeable test statistics by drawing a connection between infinite sequences of exchangeable -values and Archimedean copulae with completely monotone generators. Our theoretical results are applied to important copula families, including Clayton copulae and Gumbel copulae.
View on arXiv