Goodness-of-fit statistics for anomaly detection in Chung-Lu random graphs

Benjamin A. Miller, Lauren H. Stephens, Nadya Bliss

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

6 Citations (Scopus)

Abstract

Anomaly detection in graphs is a relevant problem in numerous applications. When determining whether an observation is anomalous with respect to the model of typical behavior, the notion of "goodness of fit" is important. This notion, however, is not well-understood in the context of graph data. In this paper, we propose three goodness-of-fit statistics for Chung-Lu random graphs, and analyze their efficacy in discriminating graphs generated by the Chung-Lu model from those with anomalous topologies. In the results of a Monte Carlo simulation, we see that the most powerful statistic for anomaly detection depends on the type of anomaly, suggesting that a hybrid statistic would be the most powerful.

Original languageEnglish (US)
Title of host publicationICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
Pages3265-3268
Number of pages4
DOIs
StatePublished - 2012
Externally publishedYes
Event2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012 - Kyoto, Japan
Duration: Mar 25 2012Mar 30 2012

Other

Other2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012
CountryJapan
CityKyoto
Period3/25/123/30/12

Fingerprint

Statistics
Topology
Monte Carlo simulation

Keywords

  • anomaly detection
  • goodness of fit
  • Graph theory
  • probabilistic models
  • signal detection theory

ASJC Scopus subject areas

  • Signal Processing
  • Software
  • Electrical and Electronic Engineering

Cite this

Miller, B. A., Stephens, L. H., & Bliss, N. (2012). Goodness-of-fit statistics for anomaly detection in Chung-Lu random graphs. In ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings (pp. 3265-3268). [6288612] https://doi.org/10.1109/ICASSP.2012.6288612

Goodness-of-fit statistics for anomaly detection in Chung-Lu random graphs. / Miller, Benjamin A.; Stephens, Lauren H.; Bliss, Nadya.

ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. 2012. p. 3265-3268 6288612.

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

Miller, BA, Stephens, LH & Bliss, N 2012, Goodness-of-fit statistics for anomaly detection in Chung-Lu random graphs. in ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings., 6288612, pp. 3265-3268, 2012 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2012, Kyoto, Japan, 3/25/12. https://doi.org/10.1109/ICASSP.2012.6288612
Miller BA, Stephens LH, Bliss N. Goodness-of-fit statistics for anomaly detection in Chung-Lu random graphs. In ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. 2012. p. 3265-3268. 6288612 https://doi.org/10.1109/ICASSP.2012.6288612
Miller, Benjamin A. ; Stephens, Lauren H. ; Bliss, Nadya. / Goodness-of-fit statistics for anomaly detection in Chung-Lu random graphs. ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. 2012. pp. 3265-3268
@inproceedings{9eda6e419e8146a899663532ed3f711c,
title = "Goodness-of-fit statistics for anomaly detection in Chung-Lu random graphs",
abstract = "Anomaly detection in graphs is a relevant problem in numerous applications. When determining whether an observation is anomalous with respect to the model of typical behavior, the notion of {"}goodness of fit{"} is important. This notion, however, is not well-understood in the context of graph data. In this paper, we propose three goodness-of-fit statistics for Chung-Lu random graphs, and analyze their efficacy in discriminating graphs generated by the Chung-Lu model from those with anomalous topologies. In the results of a Monte Carlo simulation, we see that the most powerful statistic for anomaly detection depends on the type of anomaly, suggesting that a hybrid statistic would be the most powerful.",
keywords = "anomaly detection, goodness of fit, Graph theory, probabilistic models, signal detection theory",
author = "Miller, {Benjamin A.} and Stephens, {Lauren H.} and Nadya Bliss",
year = "2012",
doi = "10.1109/ICASSP.2012.6288612",
language = "English (US)",
isbn = "9781467300469",
pages = "3265--3268",
booktitle = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",

}

TY - GEN

T1 - Goodness-of-fit statistics for anomaly detection in Chung-Lu random graphs

AU - Miller, Benjamin A.

AU - Stephens, Lauren H.

AU - Bliss, Nadya

PY - 2012

Y1 - 2012

N2 - Anomaly detection in graphs is a relevant problem in numerous applications. When determining whether an observation is anomalous with respect to the model of typical behavior, the notion of "goodness of fit" is important. This notion, however, is not well-understood in the context of graph data. In this paper, we propose three goodness-of-fit statistics for Chung-Lu random graphs, and analyze their efficacy in discriminating graphs generated by the Chung-Lu model from those with anomalous topologies. In the results of a Monte Carlo simulation, we see that the most powerful statistic for anomaly detection depends on the type of anomaly, suggesting that a hybrid statistic would be the most powerful.

AB - Anomaly detection in graphs is a relevant problem in numerous applications. When determining whether an observation is anomalous with respect to the model of typical behavior, the notion of "goodness of fit" is important. This notion, however, is not well-understood in the context of graph data. In this paper, we propose three goodness-of-fit statistics for Chung-Lu random graphs, and analyze their efficacy in discriminating graphs generated by the Chung-Lu model from those with anomalous topologies. In the results of a Monte Carlo simulation, we see that the most powerful statistic for anomaly detection depends on the type of anomaly, suggesting that a hybrid statistic would be the most powerful.

KW - anomaly detection

KW - goodness of fit

KW - Graph theory

KW - probabilistic models

KW - signal detection theory

UR - http://www.scopus.com/inward/record.url?scp=84867595594&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84867595594&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2012.6288612

DO - 10.1109/ICASSP.2012.6288612

M3 - Conference contribution

AN - SCOPUS:84867595594

SN - 9781467300469

SP - 3265

EP - 3268

BT - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

ER -