Detection, prediction and mastery of complex situations are crucial to the competitiveness of networked businesses, the efficiency of Internet of Services and dynamic distributed infrastructures in manifold domains such as finance/banking, logistics, automotive, telecommunication, e-health and life sciences. Complex Event Processing (CEP) is an emerging technology to achieve actionable, situational knowledge from huge event streams in real-time or almost close to real-time.
In many business organizations some of the important complex events cannot be used in process management because they are not detected from the workflows and decision makers cannot be informed about them. Detection of events is one of the critical factors for event-driven systems and business process management.
The current successes in business process management (BPM) and enterprise application integration (EAI) makes it possible that many organizations know a lot about their own activities.
Almost all of the business activities are logged in different log and audit systems so that all they can be used to monitor the business processes. However, the huge amounts of event information cannot be used completely in the decision making and process controlling, because the specification of event detection patterns have to done manually by humans and are highly complex.
The permanent stream of low level events in business organizations needs an intelligent real-time event processor. The detection of occurrences of complex events in the organization can be used to optimize the management of business processes. The existing event processing approaches are dealing primarily with the syntactical processing of low-level signals, constructive event database views, streams, and primitive actions. Our research on semantic complex event processing provided solutions for the fusion of background knowledge with the event streams. We provided solutions (within our research project ``Corporate Semantic Web'' for the detection of complex events based on the background knowledge
In this research, we address the problem of automated extraction of patterns for detection of complex events. The existing approaches for the pattern detection are primarily dealing with syntactical processing of event sequences to detect complex patterns only based on the sequences of event happening. As an extension of the existing approaches for pattern mining, we investigate the usage of ontological background knowledge to be able to extract complex event patterns based on the relations of patterns to the resources in the background knowledge.
Project Research Associate: Ahmad Hasan, Dr. Tara Athan, Dr. Kia Teymourian