Research Of The CEP-based Monitoring System On Abnormal Events In Securities Trading

With the economic globalization and the progressive opening to the capicalmarket, China’s securities market has obtained a big development. As a specialcommodity, the security has its own vituality and special utility, which cause a seriesof illegal behaviors in securities market. Thus, setting up a set of advanced andreal-time monitoring system on abnormal events in securities market is the premisefor the normal operation of the market.The securities market is a typical Big Data market, which has thecharacteristics of large amount of data, complex data structures, flow speed andlow value density. On the other hand, the monitoring system must be able toprocess the abnormal events in real time. Complex event processing is a newlydeveloped event processing technology, which process the events of hugequantity and high real-time like securities trading, Internet of things and medicalcare etc. Complex event processing platform has its own special eventdescription language and powerful events engine, the latter can independentlycomplete the reception, analysis and processing of large-scale data. The enginecan be integrated into the system as a component, which makes it has thecharacteristics of high flexibility and low coupling.Based on the research of complex events processing, the thesis employed c#procedural language and StreamInsight complex events processing engine to developa console desk process which monitors the suspicious securities trading events. Thesystem adopts layered structure mode, which has a clear programmed flowing moduleand a certain degree of easy maintenance, operation and expansion.Firstly, the thesis elaborated the research status and the related technologies ofbig data and complex events processing, and had a research on the related theory ofcomplex events processing.Secondly, this thesis summarized the characteristics of two suspicious events insecurities trading, which are whipsawing and knowing. They are described and stored through events algebra and XML language. Complex events modes are stored in thedatabase of rules and used as the rules of monitoring the suspicious events.Based on the LINQ transformed from the two suspicious events’ algevra andXML description, the ending case part developed a StreamInsight-based console deskprocess.By transforming the historical transaction data into data stream by means ofcomputational experiments, the thesis simulated these data stream into the eventsprocessing engine and carried out analysis on them, and finally found out thesuspicious users and behaviors，which is a good example to show the high property ofCEP in processing Big Datas in securities trading.