Speaker: Zhong Wei, professor and doctoral supervisor, Xiamen University

Date: April 18, 2025

Time: 14:30-15:30 pm

Location: B1238, Zhixin Building, Shandong University

Sponsor: School of Mathematics, Shandong University

Abstract:

Stability and reproducibility are essential considerations in various applications of statistical and machine learning methods to scientific research. False Discovery Rate (FDR) control methods provide a framework for controlling false signals in scientific discoveries. Numerous FDR control techniques, such as knockoff methods and data-splitting approaches, have been successfully developed and widely implemented in multiple testing and regression contexts. However, some of these methods can yield unstable results due to the inherent randomness of the algorithms. For instance, different constructions of knockoff copies can lead to different sets of selected variables. To enhance the stability and reproducibility of statistical outcomes, we propose a unified stability approach for feature selection and multiple testing algorithms with FDR control, named Stabilized eBH. Our method aggregates e-values based on rank statistics generated from multiple runs of the base algorithm to construct stabilized e-values, which are then processed using the eBH procedure. This approach not only improves FDR control and power performance but also enhances the stability. It is adaptable and can be applied to most existing FDR control methods. Moreover, we investigate the theoretical properties of the stability method, including asymptotic FDR control, power enhancement, and stability guarantee. Extensive numerical experiments and applications to real datasets demonstrate that the proposed method generally outperforms existing alternatives.

For more information, please visit:

https://www.view.sdu.edu.cn/info/1020/201186.htm