This paper illustrates how real-time sensor data, often described as the "event cloud," can be processed and refined using rule-based event-decision methods. From event streams, it is possible to extract features in real time, based on matching patterns in the event stream. Features of the raw data or event stream become "events of interest," which can be scored according to likelihood estimates constructed from high speed pattern matching algorithms with low latency. Complex events and situations discovered in runtime can be processed with patterns developed from historical data to predict business impact and other future events and situations.
Fraud detection is but one example of how the concepts of Predictive Business can help organizations minimize threats and liabilities to their customers - the enterprises that serve them.
