媒体信息内容与中国股市中的股票收益的关系

发布时间：2018-01-14 21:13

本文关键词：媒体信息内容与中国股市中的股票收益的关系　出处：《哈尔滨工业大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着中国股票市场的蓬勃发展，投资者对媒体信息的关注度越来越高，而媒体信息对投资者的关注度和投资策略的影响也是越来越大。因为个人投资者的关注力和认知能力的有限，他们只能选择性的处理他们所获得的全部信息里的一部分，所以相对于其他信息而言更权威的、更可信的媒体信息就成为了散户投资者们的首选，进而影响了股票收益。虽然人们认同媒体信息对投资者的行为以及上市公司股票价格、收益的影响，但系统全面的研究媒体信息到底如何影响投资者和股票收益，国内才刚刚开始，而对媒体信息内容进行文本倾向性分析，进而探讨其对上市公司股票收益的影响的研究更是少之又少，所以目前非常需要对该领域进行更为全面和深入的研究。本文首先采用篇章级的文本倾向性分析方法，对媒体信息进行文本倾向性分类，分出积极正向的文章和消极负向的文章。分类算法采用本文提出的HMSA分类算法，同时与为K-最近邻算法、最大熵分类算法和支持向量机分类算法进行比较，实验结果表明，HMSA算法的分类效果最好，准确率为79.71%，召回率达到78.00%，F值为78.85%。在文本倾向性分类后，验证了媒体信息内容对股票收益的影响机理，媒体信息首先对投资者的投资策略产生影响，进而影响了股票的价格、成交量和收益率。通过分析发现积极正向的媒体信息可以有效的预测股票收益率的上涨和随后的收益率下跌，而消极负向的媒体信息可以有效的预测股票收益率的下跌和随后的收益率上涨。在对影响机理验证结果的基础上，研究了基于媒体信息文本倾向性指标的投资策略获利模式。由于媒体信息明显地影响了股票收益，，构建了基于媒体信息文本指标的零投资组合的投资策略和不考虑该指标的零投资组合的投资策略。通过比较，发现引入该指标后可以获得较高的收益回报，增加了投资者的获利空间，对理论研究和实际投资都提供了有益的帮助。
[Abstract]:With the vigorous development of Chinese stock market, investors pay more and more attention to media information. The influence of media information on investors' attention and investment strategy is also increasing, because of the limited attention and cognitive ability of individual investors. They can only deal selectively with some of the information they get, so more authoritative and credible media information than other information is the preferred choice for retail investors. And then affect the stock returns, although people agree with the media information on the behavior of investors and listed companies stock prices, returns. However, the systematic and comprehensive research on how media information affects investors and stock returns is just beginning in China, and the text orientation analysis of media information content is carried out. Therefore, it is necessary to do more comprehensive and in-depth research in this field. Firstly, the text orientation analysis method is used to classify the media information. The classification algorithm adopts the HMSA classification algorithm proposed in this paper, and at the same time, it is the K-nearest neighbor algorithm. The maximum entropy classification algorithm is compared with the support vector machine classification algorithm. The experimental results show that HMSA algorithm has the best classification effect, the accuracy is 79.71 and the recall rate is 78.00%. F value is 78.85. After the text orientation classification, the paper verifies the influence mechanism of media information content on stock returns. Media information first affects investors' investment strategy, and then affects the stock price. The positive positive media information can effectively predict the rise of stock yield and the subsequent decline of yield. Negative media information can effectively predict the decline of stock yield and the subsequent increase in yield. Based on the results of verification of the influence mechanism, this paper studies the profit model of investment strategy based on the tendency index of media information text, because media information has a significant impact on stock returns. A zero-portfolio investment strategy based on the media information text index and a zero-portfolio investment strategy without considering the index are constructed. By comparison, it is found that the introduction of the index can achieve a higher return. It increases the profit space of investors and provides beneficial help for both theoretical research and practical investment.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.51;TP391.1

【参考文献】