Analytical Models and Methods for Anomaly Detection in Dynamic, Attributed Graphs

Benjamin A. Miller, Nicholas Arcolano, Stephen Kelley, Nadya Bliss

Research output: Chapter in Book/Report/Conference proceedingChapter

1 Citation (Scopus)

Abstract

When working with large-scale network data, the interconnected entities often have additional descriptive information. The additional metadata may provide insight that can be exploited for detection of anomalous events. This chapter provides an overview of anomaly detection techniques in dynamic, attributed graphs, with a specific focus on detection within backgrounds based on random graph models. This kind of analysis can be applied for a variety of background models, including those with dynamic topologies and where vertices and edges have attributes. A spectral framework for anomalous subgraph detection in random background models is described in detail, including an implementation in R that exploits structure in the random graph models for computationally tractable analysis of graph residuals. Detection results, both in simulation and in real data, demonstrate the power of this approach and the advantage of accounting for dynamics and attributes when analyzing graph data.

Original languageEnglish (US)
Title of host publicationComputational Network Analysis with R
Subtitle of host publicationApplications in Biology, Medicine and Chemistry
PublisherWiley-Blackwell
Pages35-61
Number of pages27
ISBN (Electronic)9783527694365
ISBN (Print)9783527339587
DOIs
StatePublished - Jul 31 2016

Fingerprint

Metadata

ASJC Scopus subject areas

  • Medicine(all)

Cite this

Miller, B. A., Arcolano, N., Kelley, S., & Bliss, N. (2016). Analytical Models and Methods for Anomaly Detection in Dynamic, Attributed Graphs. In Computational Network Analysis with R: Applications in Biology, Medicine and Chemistry (pp. 35-61). Wiley-Blackwell. https://doi.org/10.1002/9783527694365.ch2

Analytical Models and Methods for Anomaly Detection in Dynamic, Attributed Graphs. / Miller, Benjamin A.; Arcolano, Nicholas; Kelley, Stephen; Bliss, Nadya.

Computational Network Analysis with R: Applications in Biology, Medicine and Chemistry. Wiley-Blackwell, 2016. p. 35-61.

Research output: Chapter in Book/Report/Conference proceedingChapter

Miller, BA, Arcolano, N, Kelley, S & Bliss, N 2016, Analytical Models and Methods for Anomaly Detection in Dynamic, Attributed Graphs. in Computational Network Analysis with R: Applications in Biology, Medicine and Chemistry. Wiley-Blackwell, pp. 35-61. https://doi.org/10.1002/9783527694365.ch2
Miller BA, Arcolano N, Kelley S, Bliss N. Analytical Models and Methods for Anomaly Detection in Dynamic, Attributed Graphs. In Computational Network Analysis with R: Applications in Biology, Medicine and Chemistry. Wiley-Blackwell. 2016. p. 35-61 https://doi.org/10.1002/9783527694365.ch2
Miller, Benjamin A. ; Arcolano, Nicholas ; Kelley, Stephen ; Bliss, Nadya. / Analytical Models and Methods for Anomaly Detection in Dynamic, Attributed Graphs. Computational Network Analysis with R: Applications in Biology, Medicine and Chemistry. Wiley-Blackwell, 2016. pp. 35-61
@inbook{4d19f240ebfe49f8963740405f29e25d,
title = "Analytical Models and Methods for Anomaly Detection in Dynamic, Attributed Graphs",
abstract = "When working with large-scale network data, the interconnected entities often have additional descriptive information. The additional metadata may provide insight that can be exploited for detection of anomalous events. This chapter provides an overview of anomaly detection techniques in dynamic, attributed graphs, with a specific focus on detection within backgrounds based on random graph models. This kind of analysis can be applied for a variety of background models, including those with dynamic topologies and where vertices and edges have attributes. A spectral framework for anomalous subgraph detection in random background models is described in detail, including an implementation in R that exploits structure in the random graph models for computationally tractable analysis of graph residuals. Detection results, both in simulation and in real data, demonstrate the power of this approach and the advantage of accounting for dynamics and attributes when analyzing graph data.",
author = "Miller, {Benjamin A.} and Nicholas Arcolano and Stephen Kelley and Nadya Bliss",
year = "2016",
month = "7",
day = "31",
doi = "10.1002/9783527694365.ch2",
language = "English (US)",
isbn = "9783527339587",
pages = "35--61",
booktitle = "Computational Network Analysis with R",
publisher = "Wiley-Blackwell",
address = "United States",

}

TY - CHAP

T1 - Analytical Models and Methods for Anomaly Detection in Dynamic, Attributed Graphs

AU - Miller, Benjamin A.

AU - Arcolano, Nicholas

AU - Kelley, Stephen

AU - Bliss, Nadya

PY - 2016/7/31

Y1 - 2016/7/31

N2 - When working with large-scale network data, the interconnected entities often have additional descriptive information. The additional metadata may provide insight that can be exploited for detection of anomalous events. This chapter provides an overview of anomaly detection techniques in dynamic, attributed graphs, with a specific focus on detection within backgrounds based on random graph models. This kind of analysis can be applied for a variety of background models, including those with dynamic topologies and where vertices and edges have attributes. A spectral framework for anomalous subgraph detection in random background models is described in detail, including an implementation in R that exploits structure in the random graph models for computationally tractable analysis of graph residuals. Detection results, both in simulation and in real data, demonstrate the power of this approach and the advantage of accounting for dynamics and attributes when analyzing graph data.

AB - When working with large-scale network data, the interconnected entities often have additional descriptive information. The additional metadata may provide insight that can be exploited for detection of anomalous events. This chapter provides an overview of anomaly detection techniques in dynamic, attributed graphs, with a specific focus on detection within backgrounds based on random graph models. This kind of analysis can be applied for a variety of background models, including those with dynamic topologies and where vertices and edges have attributes. A spectral framework for anomalous subgraph detection in random background models is described in detail, including an implementation in R that exploits structure in the random graph models for computationally tractable analysis of graph residuals. Detection results, both in simulation and in real data, demonstrate the power of this approach and the advantage of accounting for dynamics and attributes when analyzing graph data.

UR - http://www.scopus.com/inward/record.url?scp=85046783022&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85046783022&partnerID=8YFLogxK

U2 - 10.1002/9783527694365.ch2

DO - 10.1002/9783527694365.ch2

M3 - Chapter

SN - 9783527339587

SP - 35

EP - 61

BT - Computational Network Analysis with R

PB - Wiley-Blackwell

ER -