用大数据方法研究股市问题

发布时间：2018-05-31 04:30

本文选题：大数据 + 股票　；参考：《沈阳师范大学》2017年硕士论文

【摘要】：股票市场是一个非常庞大而复杂的系统,其功能主要是对已经发行的股票进行转让、买卖和流通;其状况如何与国民经济的发展息息相关。在中国,伴随着国民经济的迅猛崛起,股票市场也在蓬勃的发展,越来越多的股民怀着支持国家建设与投资理财的热情投身于股市当中。因此,关于股市的研究具有重要的意义。虽说这方面的研究工作已经很多,但大多结果都不够令人满意,其主要不足之处大多在于以下两点:一、受到信息摄取量和储存水平的限制与计算水平的限制,用于支撑研究的数据量较小;二、没有充分地考虑利率与内部收益率的影响。当代计算机与互联网络的发展,使得我们有能力克服这两个不足之处,做出更加理想的科研成果。本文试从扩大数据量与重视利息率和收益率的视域出发建构三项探讨股市奥秘的方法并进行一定的实证分析。文章共五章。第一章介绍股票的意义与背景和大数据方法的意义与背景,以及相关的研究动态,并说明笔者所要进行的工作与文章的结构。第二章首先简介行为金融学的内容与意义,以及相关研究动态;然后从大数据、机器学习和行为金融学的角度出发,通过设计一个算法定义出一个表现按照某种初级炒股行为买卖一支股票的收益状况的随机变量R,并进行有关研究。基于这支股票的一定的折现历史数据运用所设计的算法求出R的一个样本;根据该样本做出一个经验分布;再根据该经验分布建立其近似解析分布的一个序,并进而给出一种求优化近似解析分布的方法。基于R的生成方式及相关探讨展示它的意义,并揭示出一些值得进一步研究的问题。第三章先从消除利率与内部收益率影响的视角出发,定义一种现值均线,而后建立一种根据长短现值均线预测股价与股指走势的方法,并基于过去多年的历史数据进行实证分析。第四章先简介马尔科夫链的有关基础知识,定义一种拟内部收益率,而后建立一种基于拟内部收益率折现历史数据与马尔科夫链预测股票与股指趋势的方法,并根据一定的历史数据进行实证分析。第五章是结论和展望,总结性的描述所做工作;指不足,并提出有待进一步研究的问题,及笔者对于未来的展望。所做工作可为炒股提供一定的启示,可为从大数据、机器学习和行为金融学的角度出发进一步研究股市提供一定的启示。
[Abstract]:The stock market is a very large and complex system, whose function is mainly to transfer, trade and circulate the issued stock, and how its status is closely related to the development of the national economy. In China, with the rapid rise of the national economy, the stock market is also booming, more and more investors are devoted to the stock market with the enthusiasm of supporting national construction and investment management. Therefore, the research on the stock market has important significance. Although much research has been done in this area, most of the results are not satisfactory, and the main shortcomings lie in the following two points: first, limited by the level of information intake and storage and limited by the level of calculation, The amount of data used to support the research is small; second, the influence of interest rate and internal rate of return is not fully considered. With the development of computer and Internet, we have the ability to overcome these two deficiencies and make more ideal scientific research results. This paper tries to construct three methods to probe into the mystery of stock market from the view of enlarging the amount of data and paying attention to the interest rate and yield. The article consists of five chapters. The first chapter introduces the significance and background of the stock and the significance and background of the big data method, and the related research trends, and explains the work to be carried out and the structure of the article. The second chapter introduces the content and significance of behavioral finance and the related research trends, and then from the perspective of big data, machine learning and behavioral finance, By designing an algorithm, we define a random variable R, which represents the income of buying and selling a stock according to some primary stock speculation, and carry on the relevant research. Based on the discounted historical data of this stock, a sample of R is obtained by using the algorithm designed, an empirical distribution is made according to the sample, and an order of approximate analytic distribution is established according to the empirical distribution. Furthermore, a method to optimize the approximate analytic distribution is given. Based on the R generation method and related discussion, the significance of R is demonstrated, and some problems worthy of further study are revealed. In the third chapter, from the perspective of eliminating the influence of interest rate and internal rate of return, we define a present value mean line, and then establish a method to predict the trend of stock price and stock index based on the long and long present value average. And based on the past years of historical data for empirical analysis. Chapter four introduces the basic knowledge of Markov chain, defines a kind of quasi-internal rate of return, and then establishes a method to predict the trend of stock and stock index based on discounted historical data of quasi-internal yield and Markov chain. And according to certain historical data for empirical analysis. The fifth chapter is the conclusion and prospect, the summary description of the work done, pointing to the shortcomings, and put forward the problems to be further studied, and the author's prospects for the future. The work can provide some enlightenment for stock speculation and further study of stock market from the point of view of big data machine learning and behavioral finance.
【学位授予单位】：沈阳师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.51

【参考文献】