基于CEP的证券交易异常事件监测系统研究

发布时间：2018-01-20 07:00

本文关键词： 证券交易异常复杂事件处理大数据事件代数 StreamInsight　出处：《浙江理工大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着经济全球化和资本市场的逐步开放，我国的证券市场得到了很大的发展。证券作为一种特殊的商品，由于其虚拟性和特殊的效用和需求特性，使得证券市场存在很多的违法违规行为。建立一套开发先进的、实时的、快速响应的证券市场交易异常事件监测系统，，是对证券市场监管、保证市场正常运行的前提。证券市场是个典型的“大数据”市场，具有数据量大、数据结构复杂、流动速度快、价值密度低等特点，此外，证券市场交易监测还需要对事件进行实时处理。复杂事件处理近年发展起来的一种新兴事件处理技术，证券交易处理是复杂事件处理中一种重要的应用领域。复杂事件处理平台有其专门的事件描述语言以及强大的处理引擎，而且该处理引擎能够独立完成大规模数据的接收、分析和处理，能够以组件的形式集成到系统中，具有低耦合、灵活性高等特点。本文在分析研究复杂事件处理技术的基础上，采用Visual Studio2010与微软复杂事件处理引擎StreamInsight的平台集成，开发一个面向证券交易可疑事件监测的控制台应用系统实例，实现了从交易数据流处理、复杂事件规则验证到证券交易事件判别等整个监测过程。论文首先对复杂事件处理以及大数据的研究现状和相关技术进行阐述，以及对复杂事件处理中的相关理论进行研究。其次，对证券交易中洗盘和对敲两种重要可疑行为的几种事件特征进行总结，并通过事件代数以及XML语言的方式对其进行描述和存储，作为复杂事件模式，存储于规则库中，来作为可疑事件监测规则。最后，对基于对敲和洗盘两类证券市场主要违法交易行为的代数定义和XML进行描述，并将其转化为StreamInsight应用程序中的LINQ查询语句，构建相应的StreamInsight复杂事件处理应用程序。同时采用计算实验方法，将股票池中的历史交易数据转换为数据流，模拟数据流进入应用程序，利用复杂事件处理引擎StreamInsight对其分析判断，查找出数据流中的异常行为事件。从而证实了复杂事件处理技术在处理证券市场“大数据”方面的优势。
[Abstract]:With the economic globalization and the gradual opening of the capital market, the securities market in China has been greatly developed. As a special commodity, the securities market is due to its virtual nature and special utility and demand characteristics. It makes the securities market have a lot of illegal behavior. To establish a set of advanced, real-time, rapid response securities market trading abnormal event monitoring system, is to monitor the securities market. The premise of ensuring the normal operation of the market. The securities market is a typical "big data" market, which has the characteristics of large amount of data, complex data structure, fast flow speed, low value density and so on. Stock market transaction monitoring also needs to deal with events in real time. Complex event processing is a new event processing technology developed in recent years. Securities transaction processing is an important application field in complex event processing. Complex event processing platform has its special event description language and powerful processing engine. Moreover, the engine can receive, analyze and process large-scale data independently, and can be integrated into the system in the form of components, which has the characteristics of low coupling, high flexibility and so on. Based on the analysis and research of complex event processing technology, the platform integration of Visual Studio2010 and Microsoft complex event processing engine StreamInsight is adopted in this paper. A console application system for monitoring suspicious events in securities transactions is developed, which realizes the whole monitoring process from transaction data flow processing, verification of complex event rules to discrimination of securities trading events. In this paper, the current research status and related technologies of complex event processing and big data, and the related theories of complex event processing are discussed. Secondly, it summarizes the event characteristics of two important suspicious behaviors in stock trading, and describes and stores them by means of event algebra and XML language. As a complex event pattern, it is stored in the rule base as a suspicious event monitoring rule. Finally, this paper describes the algebraic definition and XML of the main illegal trading behaviors in the securities market based on tapping and dishwashing. And turn it into a LINQ query statement in a StreamInsight application. The corresponding StreamInsight complex event processing application program is constructed, and the historical transaction data in stock pool is converted into data flow by using computational experiment method. The analog data stream enters the application program, and the complex event processing engine StreamInsight is used to analyze and judge it. The abnormal behavior events in the data flow are found, which proves the advantage of the complex event processing technology in dealing with "big data" in the securities market.
【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP274;F830.91

【参考文献】