Time Series Prediction Based on Event Driven Business Process Management
Jitka Kreslíková · 2014
In this paper we would like to focus on time series prediction for building trading systems based on unification of the seemingly antagonistic concepts of Complex Events Processing (CEP) and discrete, legacy business logic. In the realm of the high frequency market-making trading systems, the incoming stream of data from multiple asset sources and venues shall be processed primarily by a real-time CEP engine till the point when the statistical characteristics of the continuous results cross into the fat-tail distribution region, when a set of stringent risk-management business rules kick in.