Higher Confidence in Event Correlation Using Uncertainty Restrictions

Distributed cooperative systems that use event notification for communication can benefit from event correlation within the notification network. In the presence of uncertain data, however, correlation results easily become unreliable. The handling of uncertainty is therefore an important challenge for event correlation in distributed event notification systems. In this paper, we present a generic correlation model that is aware of uncertainty. We propose uncertainty constraints that event correlation can take into account and show how they can lead to higher confidence in the correlation result. We demonstrate that the application of this model allows to obtain a qualitative description of event correlation.