2016/11/09 by Gavin Lynch, Wenge Guo, Lynch, Gavin +5 · 2 citations
Biochemistry, Genetics and Molecular Biology · Computer Science · Decision Sciences · Mathematics · #62J15 #Algorithm #Artificial intelligence #Computer science #Control (management) #Data mining #Dependency (UML) #FOS: Computer and information sciences #FOS: Mathematics #False discovery rate #Gene expression and cancer classification #Mathematics #Methodology (stat.ME) #Multiple comparisons problem #Optimal Experimental Design Methods #Process (computing) #Sequence (biology) #Set (abstract data type) #Statistical Methods in Clinical Trials #Statistical hypothesis testing #Statistics #Statistics Theory (math.ST) #VLSI and Analog Circuit Testing #math.ST #msc:62J15 #stat.ME #stat.TH
paper · pdf · doi:10.48550/arxiv.1611.03146
published in arXiv (Cornell University) (Cornell University) · 32 pages, 3 figures
openalex publication_date 2016/11/09 · arxiv created 2016/11/10 · arxiv updated 2016/11/11 · openalex created_date 2022/09/20 · openalex updated_date 2026/08/06
Controlling the false discovery rate (FDR) is a powerful approach to multiple\ntesting. In many applications, the tested hypotheses have an inherent\nhierarchical structure. In this paper, we focus on the fixed sequence structure\nwhere the testing order of the hypotheses has been strictly specified in\nadvance. We are motivated to study such a structure, since it is the most basic\nof hierarchical structures, yet it is often seen in real applications such as\nstatistical process control and streaming data analysis. We first consider a\nconventional fixed sequence method that stops testing once an acceptance\noccurs, and develop such a method controlling the FDR under both arbitrary and\nnegative dependencies. The method under arbitrary dependency is shown to be\nunimprovable without losing control of the FDR and unlike existing FDR methods;\nit cannot be improved even by restricting to the usual positive regression\ndependence on subset (PRDS) condition. To account for any potential mistakes in\nthe ordering of the tests, we extend the conventional fixed sequence method to\none that allows more but a given number of acceptances. Simulation studies show\nthat the proposed procedures can be powerful alternatives to existing FDR\ncontrolling procedures. The proposed procedures are illustrated through a real\ndata set from a microarray experiment.\n