Leman Akoglu

Assistant Professor of Information Systems

Quick links

CONFERENCE          WORKSHOP         CODE          BOOK CHAP.          TECH. REPORTS          THESIS          PATENTS

Publications (Google Scholar, DBLP)



Book Chapter

Technical Reports

Code pdf-file-icon

  • StreamSpot: finding anomalies among a collection of streaming heterogeneous graphs

  • CARE: a sequential ensemble algorithm for outlier detection

  • HINside: ranking nodes in a heterogeneous network with node-location information

  • EdgeCentric: finding distributional anomalies for spotting suspicious user behavior.

  • TemporalOpinionSpam: event detection in multivariate indicative signals of opinion spam.

  • SpammerGroups: computing NFS (network footprint score) and identifying groups of spammers that target the same services within the same campaign.

  • GRASP-RLS: finding robust subgraphs of a given size (in Matlab).

  • FraudEagle: finding fake reviewers in bipartite review networks (in Matlab).

  • CoClusLSH: fast bottom-up (agglomerative) hierarchical co-clustering (in Matlab).

  • RTG: generate realistic, time-evolving, weighted graphs (in Matlab).

  • OddBall: spot anomalous nodes in weighted, unlabeled graphs (in Python and Matlab).
       >>> Here is a Python-free 'light' version of Oddball.

  • PICS: partition attributed graphs into cohesive clusters (in Matlab).

  • Dot2Dot: summarize (by a forest of connection trees) a set of nodes in a graph (in Matlab).

  • CompreX: find anomalies in categorical/numerical databases (in Matlab).

Ph.D. Thesis