Discovering Rules for Clustering and Predicting Asynchronous Events

Contents

A wide variety of complex systems, ranging from nuclear power plants to governments, generate asynchronous events. We present Multi-Event Dependency Detection (MEDD), a novel algorithm for acquiring event correlation rules from historical logs of asynchronous events. Given a new stream of events being generated in real time, the rules support two important functions: clustering sets of related events and predicting events that will occur in the future.