Contextualised Event-driven Prediction with Ontology-based Similarity

Sinan Sen, Jun Ma

Event-driven processing becomes ever important for applications such as reactive context-aware mobile applications, attention-handling and pervasive collaboration systems etc. However today's reactive systems define complex events with rather precise specifications. In some applications, such as fraud or failure detection, identification of similar event patterns may be of tremendous use. These kind of applications need to identify not only critical situations but also situations which are similar enough to them. We present a novel approach for event-driven processing which is realized by combining reactive rules with ontologies. Ontologies are used to capture the context in which certain active behavior is appropriate (i.e., to discover situations in which particular reactive rules fire). Second, ontologies together with similarity search techniques are utilised to enable discovery of similar complex event patterns.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.