Anomaly Detection Using Fusion of Graph Invariants on a Time Series of Graphs

• Activity Number: 637
• Type: Contributed
• Date/Time: Thursday, August 5, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Youngser Park*+ and Carey E. Priebe and Abdou Youssef
• Companies: The Johns Hopkins University and The Johns Hopkins University and The George Washington University
• Keywords: anomaly detection ; fusion ; feature representation ; time series analysis ; clustering ; graph theory

Abstract

It is known that fusion of information from graph features, compared to individual features, can provide superior inference for anomaly detection. We present a multivariate methodology for fusion of features derived from time series of graphs, and investigate its inferential efficacy. The results demonstrate that our methodology has higher detection, estimation, and localization power than standard linear weighting fusion techniques for certain anomaly classes. Simulation results using a latent process model for time series of graphs, as well as illustrative results from a time series of Enron email data, are presented.