Pattern Matching Algorithm in Event Graphs

Event-graphs are used for modelling and analyzing discrete data that usually represent activities and so on. In this work we take into analysis this kind of graphs and the problem of identifying correlations among events. We propose a methodology to first identify events that are significantly correlated and then to event patterns. Both are often related to individual entities and their behaviour, so that their specificity can be spotted in a complex event network. In the last part of the paper we describe our experimental results.