Efficient Outlier Detection in Hyperedge Streams Using MinHash and Locality-Sensitive Hashing
Mining outliers in graph data is a rapidly growing area of research. Traditional methods focus either on static graphs, or restrict relationships to be pairwise. In this work we address both of these limitations directly, and propose the first approach for mining outliers in hyperedge streams. Hyperedges, which generalize edges, faithfully capture higher order relationships that naturally occur in complex systems. Our model annotates every incoming hyperedge with an outlier score, which is based on the incident vertices and the historical relationships among them. Additionally, we describe an approximation scheme that ensures our model is suitable for being run in streaming environments. Experimental results on several real-world datasets show our model effectively identifies outliers, and that our approximation provides speedups between 33–775x.
This material is based on work supported in part by the Department of Energy National Nuclear Security Administration under Award Number(s) DE-NA0002576, NSF grant 1029711, the DOE SDAVI Institute, and the U.S. National Science Foundation (Expeditions in Computing program).
- 1.Enron network dataset—KONECT. http://konect.uni-koblenz.de/networks/enron. May 2015
- 2.Akoglu, L., McGlohon, M., Faloutsos, C.: Oddball: Spotting anomalies in weighted graphs. In: Advances in Knowledge Discovery and Data Mining, pp. 410–421. Springer (2010)Google Scholar
- 5.Broder, A.: Identifying and filtering near-duplicate documents. In: Combinatorial Pattern Matching, pp. 1–10. Springer (2000)Google Scholar
- 6.Broder, A.Z.: On the resemblance and containment of documents. In: Compression and Complexity of Sequences 1997. Proceedings, pp. 21–29. IEEE (1997)Google Scholar
- 8.El-Yaniv, R., Nisenson, M.: Optimal single-class classification strategies. In: Advances in Neural Information Processing Systems, pp. 377–384 (2007)Google Scholar
- 10.Leskovec, J., Rajaraman, A., Ullman, J.D.: Mining of Massive Datasets. Cambridge university press (2014)Google Scholar
- 11.Park, Y., Priebe, C., Marchette, D., Youssef, A.: Anomaly detection using scan statistics on time series hypergraphs. In: Link Analysis, Counterterrorism and Security (LACTS) Conference, p. 9 (2009)Google Scholar
- 13.Ranshous, S., Harenberg, S., Sharma, K., Samatova, N.F.: A scalable approach for outlier detection in edge streams using sketch-based approximations. In: Proceedings of the 2016 SIAM International Conference on Data Mining, pp. 189–197. SIAM (2016)Google Scholar
- 17.Sun, J., Qu, H., Chakrabarti, D., Faloutsos, C.: Neighborhood formation and anomaly detection in bipartite graphs. In: Fifth IEEE International Conference on Data Mining (ICDM’05), p. 8. IEEE (2005)Google Scholar
- 18.Wei, L., Qian, W., Zhou, A., Jin, W., Jeffrey, X.Y.: Hot: Hypergraph-based outlier test for categorical data. In: Pacific-Asia Conference on Knowledge Discovery and Data Mining, pp. 399–410. Springer (2003)Google Scholar