Collective Outlier Detection and Enumeration with Conformalized Closed Testing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Magnani, Chiara G., Sesia, Matteo, Solari, Aldo
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918506714365952
author Magnani, Chiara G.
Sesia, Matteo
Solari, Aldo
author_facet Magnani, Chiara G.
Sesia, Matteo
Solari, Aldo
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2308_05534
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Collective Outlier Detection and Enumeration with Conformalized Closed Testing
Magnani, Chiara G.
Sesia, Matteo
Solari, Aldo
Methodology
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.
title Collective Outlier Detection and Enumeration with Conformalized Closed Testing
topic Methodology
url https://arxiv.org/abs/2308.05534