A Robust Graph-Based Algorithm for Detection and Characterization of Anomalies in Noisy Multivariate Time Series

Abstract

Detection of anomalies in multivariate time series is an important data mining task with potential applications in medical diagnosis, ecosystem modeling, and network traffic monitoring. In this paper, we present a robust graph-based algorithm for detecting anomalies in noisy multivariate time series data. A key feature of the algorithm is the alignment of kernel matrices constructed from the time series. The aligned kernel enables the algorithm to capture the dependence relationship between different time series and to support the discovery of different types of anomalies (including subsequence-based and local anomalies). We have performed extensive experiments to demonstrate the effectiveness of the proposed algorithm. We also present a case study that shows the utility of applying our algorithm to detect ecosystem disturbances in Earth science data.