Complex Event Refinement by Statistical Augmentation Model

The uncertainty of decision making in event hierarchies of CEP can be due to unreliable data sources, lack of conformance that the event which is reported has actually occurred. Also the Complex Event models which are used to define complex events are inaccurate. When the uncertain event is used for deriving complex event, it propagates its uncertainty to a higher level of event hierarchy and causes uncertainty in reasoning. This paper proposes an event refinement model based on statistical approach to augment the events to minimize the error due to uncertainty for better decision making. The proposed augmented CEP (a-CEP) is found to perform better in terms of reduction in false alarm for continuous monitoring of patient in a remote health care application. The proposed model is implemented on Drools Fusion CEP Engine using Java and it is found that the proposed a-CEP gives better results in terms of accuracy.