arrow
Return

Sequential Feature Explanations for Anomaly Detection

delete2019-01-09
delete28
delete
OA
AI
A
Amran Siddiqui *
A
Alan Fern
T
Thomas G. Dietterich
W
Weng‐Keen Wong
DOI:10.1145/3230666delete
deleteOriginal
deleteShare
deleteSave
View PDF
Abstract

Abstract

En 中文
In many applications, an anomaly detection system presents the most anomalous data instance to a human analyst, who then must determine whether the instance is truly of interest (e.g., a threat in a security setting). Unfortunately, most anomaly detectors provide no explanation about why an instance was considered anomalous, leaving the analyst with no guidance about where to begin the investigation. To address this issue, we study the problems of computing and evaluating sequential feature explanations (SFEs) for anomaly detectors. An SFE of an anomaly is a sequence of features, which are presented to the analyst one at a time (in order) until the information contained in the highlighted features is enough for the analyst to make a confident judgement about the anomaly. Since analyst effort is related to the amount of information that they consider in an investigation, an explanation's quality is related to the number of features that must be revealed to attain confidence. In this article, we first formulate the problem of optimizing SFEs for a particular density-based anomaly detector. We then present both greedy algorithms and an optimal algorithm, based on branch-and-bound search, for optimizing SFEs. Finally, we provide a large scale quantitative evaluation of these algorithms using a novel framework for evaluating explanations. The results show that our algorithms are quite effective and that our best greedy algorithm is competitive with optimal solutions.
Keywords:
Anomaly detection
anomaly explanation
anomaly interpretation
explanation evaluation
AI Summary

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.

Journal

ACM Transactions on Knowledge Discovery from Data cover
ACM Transactions on Knowledge Discovery from Data
IF:
4.8
Papers:
1.3K
Citations:
4.4K

Organization

O
Oregon State University
Scholars:
1.7W
Papers: 1.5W
Citations: 2.4W
Cited Papers

Cited Papers

Springs Of Action
err
IF0
err2023-10-31
err0
PREAI
errAlfred R Mele
errShare
errSave
Lowest- Q2 measurement of the γp → Δ reaction: Probing the pionic contribution
err2006-12-22
err0
errOAAI
errS. Stave; M. O. Distler; I. Nakagawa; N. Sparveris; P. Achenbach; C. Ayerbe Gayoso; D. Baumann; J. Bernauer; A. M. Bernstein; R. Böhm; D. Bosnar; T. Botto; A. Christopoulou; D. Dale; M. Ding; L. Doria; J. Friedrich; A. Karabarbounis; M. Makek; H. Merkel; U. Müller; R. Neuhausen; L. Nungesser; C. N. Papanicolas; A. Piegsa; J. Pochodzalla; M. Potokar; M. Seimetz; S. Širca; S. Stiliaris; Th. Walcher; M. Weis
errShare
errSave
Protococcidian Eleutheroschizon duboscqi, an Unusual Apicomplexan Interconnecting Gregarines and Cryptosporidia
err2015-04-27
err0
errOAAI
errAndrea Valigurová; Gita G. Paskerova; Andrei Diakin; Magdaléna Kováčiková; Timur G. Simdyanov
errShare
errSave
errShare
errSave
Photodetachment of Electrons from Major Negative Ions in the Lower D Region
err1972-07-01
err0
PREAI
errL. THOMAS; P. M. GONDHALEKAR; M. R. BOWMAN
errShare
errSave
errShare
errSave
no more