Prediction based on hybrid method in complex event processing

This paper presents a procedure for performing predictive analysis of complex events occurrence in time-critical complex event processing systems. Complex event processing (CEP) is one of the three basic styles of event processing, which involve the detection and identification of dependencies in the sequences of selected events from various sources. CEP allows applications to monitor several streams of events, analyze them according to predetermined rules, and respond to the identified opportunities and threats in real time. With CEP it is possible to identify and apply business intelligence over the streams of events, and it allows user to easily identify the complex events with the temporal and spatial constraints. CEP is a technology that is critical in an environment where time plays an important role, in real time decision making. CEP increases the visibility and availability of the information, making it better and faster response to the emergence of situations difficult to predict. This paper deals with the analysis and use of available techniques for classification and prediction in complex event processing. In this paper we present the verification of selected methods and their application in predicting the complex events. For this purpose was designed and implemented application CepPredictiveAnalysis. The prediction is performed by the proposed hybrid method. Despite certain limitations, the system gives acceptable accuracy in financially-oriented applications.