Collective Outlier Detection and Enumeration with Conformalized Closed Testing

TL;DR

This paper introduces a flexible, distribution-free method for collective outlier detection and enumeration that leverages conformal inference and adaptive testing, effective even with sparse or elusive signals.

Contribution

It develops a novel, principled approach for automatically selecting classifiers and testing procedures for outlier detection, integrating conformal inference with classical statistical ideas.

Findings

Effective outlier detection in high-energy physics data

Automatic classifier and test selection improves detection power

Method demonstrates strong empirical performance

Abstract

This paper develops a flexible distribution-free method for collective outlier detection and enumeration, designed for situations in which the presence of outliers can be detected powerfully even though their precise identification may be challenging due to the sparsity, weakness, or elusiveness of their signals. This method builds upon recent developments in conformal inference and integrates classical ideas from other areas, including multiple testing, locally most powerful and adaptive rank tests, and non-parametric large-sample asymptotics. The key innovation lies in developing a principled and effective approach for automatically choosing the most appropriate machine learning classifier and two-sample testing procedure for a given data set. The performance of our method is investigated through extensive empirical demonstrations, including an analysis of the LHCO high-energy particle collision data set.