A general interactive framework for false discovery rate control under structural constraints
A general interactive framework for false discovery rate control under structural constraints
复制标题
DOI：
10.1093/biomet/asaa064
复制
发表时间：
2021-06-01
期刊：
影响因子：
2.7
通讯作者：
Fithian, William
中科院分区：
文献类型：
--
作者：
Lei, Lihua;Ramdas, Aaditya;Fithian, William
我们提出了一个通用框架的基础上选择性遍历的累积规则的交互式多重测试与通用结构约束的拒绝集。它将有序多重测试的累积测试与选择后推理的数据雕刻思想相结合，允许高度灵活地适应通用结构信息。我们的程序定义了一个交互式协议，逐步修剪候选人拒绝集，开始与所有假设的集合，并缩小了每一步。通过限制在每一步的信息通过一种技术，我们称之为屏蔽，我们的协议使互动，同时控制错误的发现率在有限的样本中的任何数据自适应更新规则，分析师可以选择。我们建议更新规则的各种应用程序具有复杂的结构约束，证明选择性遍历的积累规则执行良好的问题，从凸区域检测到错误发现率控制有向无环图，并显示如何扩展的框架，以回归问题，其中敲除统计是可用的，而不是p值。
We propose a general framework based on selectively traversed accumulation rules for interactive multiple testing with generic structural constraints on the rejection set. It combines accumulation tests from ordered multiple testing with data-carving ideas from post-selection inference, allowing highly flexible adaptation to generic structural information. Our procedure defines an interactive protocol for gradually pruning a candidate rejection set, beginning with the set of all hypotheses and shrinking the set with each step. By restricting the information at each step via a technique we call masking, our protocol enables interaction while controlling the false discovery rate in finite samples for any data-adaptive update rule that the analyst may choose. We suggest update rules for a variety of applications with complex structural constraints, demonstrate that selectively traversed accumulation rules perform well in problems ranging from convex region detection to false discovery rate control on directed acyclic graphs, and show how to extend the framework to regression problems where knockoff statistics are available in lieu of p-values.