Event Mining Applications

can be defined as a system, from a single component in a jet engine manufacturing line to ubiquitous computing. As the system operates, events can be extracted and subsequently mined through a variety of methods. Unwanted events can be filtered, anomaly events can be detected, complex event processing (CEP) can be applied to aggregate and monitor the occurrence of correspondence between multiple events, significant patterns of multiple events can be extracted, and so on. To understand complex systems and deal with complex patterns, logging sys­ tems’ events with only an ID, name, and timestamp is not enough. As mentioned in Chapter 2, events need to be stored as a complex object with all their properties and relationships rather than a relational tuple. Event mining algorithms not only need to support symbolic representation of events but also complex relationships such as causality, where one event is the root cause of another. In this chapter, we discuss how event mining can be applied in different domains and overview general requirements and high-level workflow in each domain. Event Mining Applications