导　　师： 任江涛

学科专业： 081202

授予学位： 硕士

作　　者： ;

机构地区： 中山大学

摘　　要： 现代的券商公司通过开发各种功能齐备的股票行情系统,给予用户各种实时消息,从而为用户提供更全面的投资买卖判断的凭借,大大的节约了用户的使用时间。但是目前对股票进行技术分析或者最新财务分析的选股,没有对股票历史财务信息进行横向历史选股和对各个公司的当天公告进行筛选和查询。而这对于投资者投资分析有着重要的意义,实现这两个功能能够在投资分析领域为投资者节省分析时间。本文以“便利性”和“专业化”为设计理念和实现依据,设计实现了SIAS/(Stock Information Analysis System/)系统,即“股票信息分析系统”,此系统通过使用文本挖掘和编译原理的理念和知识,实现了股票信息获取和处理这两个功能。 本文通过分析股票波动信息和公告独有的文本特性,结合文本挖掘和基于编译原理的脚本语言设计的一些理念和知识,以解决上述的两个问题为目标,重点集中在“信息的获取”、“价值信息的形成”、“阅读”和“选股”这四个方面设计及实现,力图创建一个能给用户带来更便利、更专业的股票信息分析系统SIAS。由上述的四个方面,构建了系统的三个模块架构。第一部分是信息源,主要是数据和公告信息内容的获取和存储;第二部分是“价值信息获取”层,主要是实现技术分析和证券管理学中运用于波动数据的常用公式、技术指标和统计功能,为综合选股功能提供基础;使用中科院的分词系统在公告中关键字信息搜索,编写关键字权重设置脚本,实现对公告的归类和排序,并通过结构化公告内容,实现条件过滤和显示等功能,为公告摘要阅读提供基础;第三部分是用户层,通过实现策略书写功能,调用文中所述的功能函数,实现一个为用户提供书写选股策略和阅读策略平台,并最终实现用户的选股和阅读。 实验结果证明,采用了文本挖� Nowadays brokerage companies provide a variety of real-time information to users through kinds of full-featured stock trading systems, as a more comprehensive basis for judging the investment transactions, which can greatly reduce the time of investment decisions making for users. However its shortage results in making only the current technical and financial analysis on stocks selection, which make neither the horizontal stock selection, nor the selection and search on current announcements of companies every day. It is important to realize these two functions to enhance the analysis efficient for investors. Base on the idea of Convenience and Professional,we advance a system called SIAS, Stock Information Analysis System to realize the acquisition and processing of stock information by using text mining and compiler theory. In this paper, we analyze the information of stock volatility and the unique characteristics of text in the announcements by using concept and knowledge of scripts design based on compiler theory and text mining to improve the modules mentioned above. And we focus on four aspects of 'Information Access',“Formation of Valuable Information ', 'Reading' and 'Selection', trying to create a more convenient and professional SIAS system for users. Therefore we build three function modules in SIAS. The first one,“Information Source”, is mainly access and storage of announcement texts. The second one,“Access of Valuable Information”, realizes formulas, technical indicators and statistical functions applied to fluctuations data in technical analysis and portfolio management. We extract financial data by reading announcements to serve the stock selection function. And the CAS system is used to find keywords in announcements, set keywords’weight, fetch information in forms, and provide summary reading of announcements by writing scripts. The last one is the User layer, which provides a platform of reading and writing strategies by calling functions in SIAS. The results show that SIAS

关 键 词： 文本挖掘 系统 价值信息获取 阅读 选股

领　　域： [经济管理—金融学] [经济管理—国民经济]