Graph Anomaly Detection Based on Steiner Connectivity and Density

被引:16
作者
Cadena, Jose [1 ,2 ,3 ]
Chen, Feng [4 ]
Vullikanti, Anil [1 ,2 ]
机构
[1] Virginia Tech, Dept Comp Sci, Blacksburg, VA 24060 USA
[2] Virginia Tech, Biocomplex Inst, Blacksburg, VA 24060 USA
[3] Lawrence Livermore Natl Lab, Machine Learning Grp, Livermore, CA 94550 USA
[4] SUNY Albany, Dept Comp Sci, Albany, NY 12222 USA
基金
美国国家科学基金会;
关键词
Approximation algorithms; dense subgraph mining; graph anomaly detection; graph mining; parameterized complexity; scan statistics; HIGHER CRITICISM; SUBSET SCAN; ALGORITHMS; EVENTS;
D O I
10.1109/JPROC.2018.2813311
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Detecting "hotspots" and "anomalies" is a recurring problem with a wide range of applications, such as social network analysis, epidemiology, finance, and biosurveillance, among others. Networks are a common abstraction in these applications for representing complex relationships. Typically, these networks are dynamic-, i.e., they evolve over time. A number of methods have been proposed for anomaly detection in such dynamic network data sets, which are primarily based on changes in network properties. We provide a survey of the various formulations of anomaly detection in dynamic networks with a focus on "window-based" methods. Window-based methods first define a time window of past network snapshots to model normal behavior and then mark a snapshot as anomalous if it has significantly different patterns from those observed in the time window. We describe two classes of techniques: 1) generalizations of Steiner connectivity; and 2) dense subgraph mining. Both have been used extensively in window-based graph anomaly detection. We summarize the key problem formulations that have been studied using these approaches, and we describe details of some of the main techniques.
引用
收藏
页码:829 / 845
页数:17
相关论文
共 104 条
[1]   Massive quasi-clique detection [J].
Abello, J ;
Resende, MGC ;
Sudarsky, S .
LATIN 2002: THEORETICAL INFORMATICS, 2002, 2286 :598-612
[2]  
Aggarwal CC, 2011, PROC INT CONF DATA, P399, DOI 10.1109/ICDE.2011.5767885
[3]   Graph based anomaly detection and description: a survey [J].
Akoglu, Leman ;
Tong, Hanghang ;
Koutra, Danai .
DATA MINING AND KNOWLEDGE DISCOVERY, 2015, 29 (03) :626-688
[4]  
Akoglu L, 2010, LECT NOTES ARTIF INT, V6119, P410
[5]   COLOR-CODING [J].
ALON, N ;
YUSTER, R ;
ZWICK, U .
JOURNAL OF THE ASSOCIATION FOR COMPUTING MACHINERY, 1995, 42 (04) :844-856
[6]   Accuracy Evaluation of the Unified P-Value from Combining Correlated P-Values [J].
Alves, Gelio ;
Yu, Yi-Kuo .
PLOS ONE, 2014, 9 (03)
[7]   Non-recurrent traffic congestion detection on heterogeneous urban road networks [J].
Anbaroglu, Berk ;
Cheng, Tao ;
Heydecker, Benjamin .
TRANSPORTMETRICA A-TRANSPORT SCIENCE, 2015, 11 (09) :754-771
[8]  
Andersen R, 2009, LECT NOTES COMPUT SC, V5427, P25, DOI 10.1007/978-3-540-95995-3_3
[9]  
[Anonymous], 2012, PARAMETERIZED COMPLE
[10]  
[Anonymous], ICML