Anomaly Detection in Time Series of Graphs using Fusion of Graph Invariants

作者：Park Youngser^*; Priebe Carey E; Youssef Abdou

来源：IEEE Journal of Selected Topics in Signal Processing, 2013, 7(1): 67-75.

DOI：10.1109/JSTSP.2012.2233712

• Change point detection
• fusion
• graph invariants
• hypothesis testing
• random graphs
• statistical inference on graphs
• time series analysis

Given a time series of graphs G(t) = (V, E(t)), t = 1, 2, ... , where the fixed vertex set V represents %26quot;actors%26quot; and an edge between vertex u and vertex v at time t(uv is an element of E(t)) represents the existence of a communications event between actors u and v during the tth time period, we wish to detect anomalies and/or change points. We consider a collection of graph features, or invariants, and demonstrate that adaptive fusion provides superior inferential efficacy compared to naive equal weighting for a certain class of anomaly detection problems. Simulation results using a latent process model for time series of graphs, as well as illustrative experimental results for a time series of graphs derived from the Enron email data, show that a fusion statistic can provide superior inference compared to individual invariants alone. These results also demonstrate that an adaptive weighting scheme for fusion of invariants performs better than naive equal weighting.

• 出版日期2013-2