Adaptive novelty detection with false discovery rate guarantee

DOI10.1214/23-AOS2338arXiv2208.06685MaRDI QIDQ6151968FDOQ6151968

Ariane Marandon, Lihua Lei, Étienne Roquain, David Mary

Publication date: 11 March 2024

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: Classical false discovery rate (FDR) controlling procedures offer strong and interpretable guarantees but often lack flexibility to work with complex data. By contrast, machine learning-based classification algorithms have superior performances on modern datasets but typically fall short of error-controlling guarantees. In this paper, we make these two meet by introducing a new adaptive novelty detection procedure with FDR control, called AdaDetect. It extends the scope of recent works of multiple testing literature to the high dimensional setting, notably the one in Yang et al. (2021). We prove that AdaDetect comes with finite sample guarantees: it controls the FDR strongly and approximates the oracle in terms of the power, with explicit remainder terms that are small under mild conditions. In practice, AdaDetect can be used in combination with any machine learning-based classifier, which allows the user to choose the most relevant classification approach. We illustrate this with classical real-world datasets, for which random forest and neural network classifiers are particularly efficient. The versatility of our method is also shown with an astrophysical application.

Full work available at URL: https://arxiv.org/abs/2208.06685

zbMATH Keywords

classificationfalse discovery ratemachine learningneural networknovelty detectionknockoffadaptive multiple testingconformal \(p\)-values

Mathematics Subject Classification ID

Nonparametric hypothesis testing (62G10) Classification and discrimination; cluster analysis (statistical aspects) (62H30) Paired and multiple comparisons; multiple testing (62J15)

Cites Work

• AdaPT: An Interactive Procedure for Multiple Testing with Side Information
• Covariate-Assisted Ranking and Screening for Large-Scale Two-Sample Inference
• Title not available (Why is that?)
• The control of the false discovery rate in multiple testing under dependency.
• Large-scale multiple testing under dependence
• Simultaneous Testing of Grouped Hypotheses: Finding Needles in Multiple Haystacks
• Adaptive linear step-up procedures that control the false discovery rate
• Predictive inference with the jackknife+
• Cross-conformal predictors
• Title not available (Why is that?)
• Strong Control, Conservative Point Estimation and Simultaneous Conservative Consistency of False Discovery Rates: A Unified Approach
• Title not available (Why is that?)
• Large-Scale Simultaneous Hypothesis Testing
• Robust inference with knockoffs
• Exact calculations for false discovery proportion with application to least favorable configura\-tions
• Adaptive false discovery rate control under independence and dependence
• Oracle and Adaptive Compound Decision Rules for False Discovery Rate Control
• Title not available (Why is that?)
• Empirical Bayes Analysis of a Microarray Experiment
• On the Benjamini-Hochberg method
• Controlling the false discovery rate via knockoffs
• False discovery rate control via debiased Lasso
• Discovering the False Discovery Rate
• The p-filter: Multilayer False Discovery Rate Control for Grouped Hypotheses
• SLOPE-adaptive variable selection via convex optimization
• Controlling the number of false discoveries: application to high-dimensional genomic data
• Exact and Approximate Stepdown Methods for Multiple Hypothesis Testing
• Microarrays, empirical Bayes and the two-groups model
• Estimating the support of a high-dimensional distribution
• Fast learning rates for plug-in classifiers
• Semi-supervised novelty detection
• Density ratio estimation in machine learning. Foreword by Thomas G. Dietterich
• Title not available (Why is that?)
• On methods controlling the false discovery rate
• A Neyman–Pearson Approach to Statistical Learning
• Empirical Bayes Estimates for Large-Scale Prediction Problems
• Learning from positive and unlabeled data: a survey
• Introduction to High-Dimensional Statistics
• Doing thousands of hypothesis tests at the same time
• Title not available (Why is that?)
• Multiple testing for exploratory research
• Testing for outliers with conformal p-values
• A unified treatment of multiple testing with prior knowledge using the p-filter
• Semi-supervised multiple testing
• Conditional calibration for false discovery rate control under dependence
• Convergence rates of deep ReLU networks for multiclass classification
• Global and Simultaneous Hypothesis Testing for High-Dimensional Logistic Regression Models
• A sequential algorithm for false discovery rate control on directed acyclic graphs
• Smoothed nested testing on directed acyclic graphs

This page was built for publication: Adaptive novelty detection with false discovery rate guarantee

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6151968)