An approach to modeling and discovering event correlation for service collaboration

Abstract

In an IoT (Internet of Things) environment, event correlation becomes more complex as events usually span over many interrelated sensors. This paper refines event correlations in an IoT environment. We extend our previous service hyperlink model to encapsulate such event correlations. To effectively discover service hyperlinks, we transform the event correlation discovery problem into a frequent sequence mining problem and propose CorFinder algorithm. Moreover, we apply our approach to improve anomaly warning in a power plant instead of simulation. Besides the application, we have made extensive experiments to verify the effectiveness of our approach.