Abstract

Anomaly detection in data streams requires a signal of an unusual event, but an actionable response requires diagnostics. Consequently, an important task is to isolate to the few key attributes that contribute to the signal from among a large collection. We introduce this contributor problem to the machine learning community and present a solution for monitoring in modern systems (with nonlinear reference conditions, high dimensions, categorical attributes, missing data, and so forth). The objective is to identify attributes that contribute to a signal, for both individual and multiple anomalies, or from several anomaly groups. Although related to the feature selection problem, the extreme sparseness of anomalies leads to scores that are designed specifically for the contributors problem. Statistical criteria are provided to quantitatively address decision rules and false alarms and the method can be computed quickly. Comparisons are made to traditional contribution plots.

Original languageEnglish (US)
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages944-953
Number of pages10
Volume5769 LNCS
EditionPART 2
DOIs
StatePublished - 2009
Event19th International Conference on Artificial Neural Networks, ICANN 2009 - Limassol, Cyprus
Duration: Sep 14 2009Sep 17 2009

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume5769 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other19th International Conference on Artificial Neural Networks, ICANN 2009
CountryCyprus
CityLimassol
Period9/14/099/17/09

Fingerprint

Anomaly Detection
Anomaly
Learning systems
Feature extraction
Diagnostics
Attribute
Monitoring
False Alarm
Decision Rules
Missing Data
Data Streams
Categorical
Feature Selection
Higher Dimensions
Machine Learning
Extremes

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Borisov, A., Runger, G., & Tuv, E. (2009). Contributor diagnostics for anomaly detection. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 2 ed., Vol. 5769 LNCS, pp. 944-953). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5769 LNCS, No. PART 2). https://doi.org/10.1007/978-3-642-04277-5_95

Contributor diagnostics for anomaly detection. / Borisov, Alexander; Runger, George; Tuv, Eugene.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5769 LNCS PART 2. ed. 2009. p. 944-953 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5769 LNCS, No. PART 2).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Borisov, A, Runger, G & Tuv, E 2009, Contributor diagnostics for anomaly detection. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 2 edn, vol. 5769 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 5769 LNCS, pp. 944-953, 19th International Conference on Artificial Neural Networks, ICANN 2009, Limassol, Cyprus, 9/14/09. https://doi.org/10.1007/978-3-642-04277-5_95
Borisov A, Runger G, Tuv E. Contributor diagnostics for anomaly detection. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 2 ed. Vol. 5769 LNCS. 2009. p. 944-953. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). https://doi.org/10.1007/978-3-642-04277-5_95
Borisov, Alexander ; Runger, George ; Tuv, Eugene. / Contributor diagnostics for anomaly detection. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5769 LNCS PART 2. ed. 2009. pp. 944-953 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).
@inproceedings{8ec7b817c62146d9861fa746a77f68bf,
title = "Contributor diagnostics for anomaly detection",
abstract = "Anomaly detection in data streams requires a signal of an unusual event, but an actionable response requires diagnostics. Consequently, an important task is to isolate to the few key attributes that contribute to the signal from among a large collection. We introduce this contributor problem to the machine learning community and present a solution for monitoring in modern systems (with nonlinear reference conditions, high dimensions, categorical attributes, missing data, and so forth). The objective is to identify attributes that contribute to a signal, for both individual and multiple anomalies, or from several anomaly groups. Although related to the feature selection problem, the extreme sparseness of anomalies leads to scores that are designed specifically for the contributors problem. Statistical criteria are provided to quantitatively address decision rules and false alarms and the method can be computed quickly. Comparisons are made to traditional contribution plots.",
author = "Alexander Borisov and George Runger and Eugene Tuv",
year = "2009",
doi = "10.1007/978-3-642-04277-5_95",
language = "English (US)",
isbn = "3642042767",
volume = "5769 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 2",
pages = "944--953",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
edition = "PART 2",

}

TY - GEN

T1 - Contributor diagnostics for anomaly detection

AU - Borisov, Alexander

AU - Runger, George

AU - Tuv, Eugene

PY - 2009

Y1 - 2009

N2 - Anomaly detection in data streams requires a signal of an unusual event, but an actionable response requires diagnostics. Consequently, an important task is to isolate to the few key attributes that contribute to the signal from among a large collection. We introduce this contributor problem to the machine learning community and present a solution for monitoring in modern systems (with nonlinear reference conditions, high dimensions, categorical attributes, missing data, and so forth). The objective is to identify attributes that contribute to a signal, for both individual and multiple anomalies, or from several anomaly groups. Although related to the feature selection problem, the extreme sparseness of anomalies leads to scores that are designed specifically for the contributors problem. Statistical criteria are provided to quantitatively address decision rules and false alarms and the method can be computed quickly. Comparisons are made to traditional contribution plots.

AB - Anomaly detection in data streams requires a signal of an unusual event, but an actionable response requires diagnostics. Consequently, an important task is to isolate to the few key attributes that contribute to the signal from among a large collection. We introduce this contributor problem to the machine learning community and present a solution for monitoring in modern systems (with nonlinear reference conditions, high dimensions, categorical attributes, missing data, and so forth). The objective is to identify attributes that contribute to a signal, for both individual and multiple anomalies, or from several anomaly groups. Although related to the feature selection problem, the extreme sparseness of anomalies leads to scores that are designed specifically for the contributors problem. Statistical criteria are provided to quantitatively address decision rules and false alarms and the method can be computed quickly. Comparisons are made to traditional contribution plots.

UR - http://www.scopus.com/inward/record.url?scp=70450204281&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70450204281&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-04277-5_95

DO - 10.1007/978-3-642-04277-5_95

M3 - Conference contribution

AN - SCOPUS:70450204281

SN - 3642042767

SN - 9783642042768

VL - 5769 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 944

EP - 953

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -