A Domain-Driven Approach for Detecting Event Patterns in E-Markets: A Case Study in Financial Market Surveillance

摘要

An e-market can be thought of as a distributed event system where an event is generated every time the market's state changes in response to a number of human or computing agents. The paper describes a practical application of event processing in an e-market context through conventional Event Processing Systems (EPSs). A new EPS architecture that allows an integration of several existing EPSs under a unified domain-specific user interface and execution environment is proposed. We assess the performance of the system for a case study in financial market surveillance and its ability to provide a common interface for two existing EPSs - SMARTS and Coral8. A discussion on the experimental results and the issues arising from the proposed EPS architecture are also provided.