基于大数据挖掘的TMT行业情绪指数的编写

发布时间：2018-06-12 19:29

本文选题：文本挖掘 + 量化投资　；参考：《上海交通大学》2015年硕士论文

【摘要】：互联网应用的渗透不断上升使得网上的信息量爆发性增长,与此同时,通过对于互联网上的文本挖掘,产生了不同于原有投资分析方法的新的框架。应用市场情绪指标的投资策略应运而生。本文首先研究对比了传统行业的分析框架以及新兴行业的分析框架,并对现有的市场情绪分析方法就行了梳理和总结。其次,分析和研究爬虫算法以及网页数据结构,设计了网络爬虫程序,对互联网上的用户讨论信息进行了抓取,为指数的构建提供了数据基础。再次,通过设定主题讨论数量、主流媒体新闻报道量、资金流入占比以及板块内涨停版数目四个参数作为情绪指标的代理变量,对其进行处理和合成,构建了情绪指标的模型。并以此设计了投资策略。最后运用构建完成的指标,对智能家居以及3D打印两个TMT行业的板块进行了实证研究。验证了情绪指标在投资决策过程中的实践意义。
[Abstract]:The penetration of Internet applications has increased the amount of information on the Internet. At the same time, through the text mining on the Internet, a new framework which is different from the original method of investment analysis has been created. The analysis framework of the emerging industry and the existing market sentiment analysis methods have been combed and summarized. Secondly, the crawler algorithm and the web data structure are analyzed and studied, the web crawler program is designed, the user discussion information on the Internet is captured, and the data base for the construction of the index is provided. Again, the theme is set by setting the subject. On the quantity, the mainstream media news coverage, the capital inflow account ratio and the four parameters of the intra plate fluctuation version as the agent variables of the emotional index, they are processed and synthesized, and the model of the emotional index is constructed. And the investment strategy is designed. Finally, the construction targets are used to print two TMT lines to the smart home and 3D. An empirical study of the industry sector is carried out to verify the practical significance of emotional indicators in the investment decision-making process.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP311.13;F832.51

【参考文献】