一种分阶段复合的语义复杂事件检测机制

复杂事件处理（CEP）是目前用于事件流处理的最佳技术,为进一步提高CEP的检测能力,检出具有更高语义抽象层次的复杂事件.本文对现存的两类检测机制进行了有效融合,利用非逻辑检测机制对原子简单事件序列时序状态进行匹配,产生具有语义的事件,而后利用逻辑检测机制推理出含有高级概念含义的复杂事件.同时给出了能够实现该机制的系统结构模型,能够进行事件模式定义的类（Structured Query Language,SQL）声明式检测语言,以及系统各组成模块所采用的关键技术.最后针对天气预测场景的仿真实验结果表明,该方法能够实时高效的检测出语义目标事件,具有一定的可行性.

A Staged Composite Complex Event Detection Mechanism

Complex event proecessing is a promising technology which can effectively handle high velocity event streams.For the purpose of enhancing the detection capability of CEP and inspecting out more complex event with high abstract level,the paper proposes an integrated detection mechanism consisting of two different detection methods.The machanism is to use the utilzes non-logical detection method to evaluate sequential state on the event stream to produce some junior semantic events,and to use the logical detection method to deduce the event which contains senior conception.The system structure model which can achieve the proposed machanism,the event pattern definition language which is declarative and SQL-like,and the key technologies used in the system are also introduced in this paper.The experimental results obtained from the the simulated weather forecasting indicate that the method can detect the target semantic event effectively in real-time,showing that the proposed machanism is feasible.