股票量化交易策略的研究及MATLAB的实现

发布时间：2018-06-03 17:58

本文选题：选股策略 + 多因子模型　；参考：《天津商业大学》2017年硕士论文

【摘要】：2008年金融危机之后,全世界金融机构均受到极大影响,国际国内的资本市场都经历了一场浩劫,几乎所有的基金、股票等金融产品都出现了不同程度的亏损,股民们也都遭遇了重大财产损失。价值投资和技术趋势投资在这一场金融风暴的影响下也失去了它应有的作用,而在此时量化投资逐渐显露出头角,在资本市场中维持了较好的收益率。的确,随着科技的不断发展发展,人工智能技术愈加成熟,为量化投资者将数理金融学、应用统计学等大数据理论与计算机相结合,以及利用计算机远远优异于人脑的计算速度,来寻找能够在整个金融市场中取得超额盈利的策略奠定了基础。在投资产品日益增多的今天,随着我国证券市场的逐渐完善,使投资者的观念也从传统的的技术分析为主导慢慢向量化投资产品转化,量化投资已成必然趋势。本文利用MATLAB软件,从WIND机构版获取股票的基本面和技术面数据,将部分上市比较短的公司剔除,以整个A股市场股票为样本进行量化分析,最后将收益率作为衡量的标准给出合理的量化投资策略。在文中,首先介绍了量化投资的相关概念和量化交易的国内、外发展现状,并对文献进行综述;其次,详细阐述了多因子选股过程以及其理论支撑;第三,选出在回测中收益率较好的因子建立了阿尔法多因子选股模型:从估值性、成长性、技术面等角度选取了15常用的因子指标作为待选因子,选取2007年1月到2013年12月为样本内检验期,对这些因子进行有效性和冗余性检验,通过比较投资组合的收益和同时间段内中证500的收益,最后得出三个有效的选股因子,并结合统计检验的方法,建立一个综合评分多因子量化选股模型,为投资者找出一个合理的符合中国市场的股票投资组合奠定了基础;最后,搭建了基于MATLAB环境下的股票量化交易平台,并利用简单策略进行股票的买入卖出功能对平台的稳定性和可操作性进行了测试。平台运行时,进行回测的历史数据来源于WIND机构版,实时交易数据由新浪股票行情客户端获得,最终通过股票API外挂到同花顺上实现股票的全自动化交易。该平台主要功能包括:股票账号登录,买入,卖出,查询资金,查询持仓,查询成交,查询委托,撤单等等。
[Abstract]:After the 2008 financial crisis, financial institutions all over the world were greatly affected. The international and domestic capital markets all experienced a catastrophe. Almost all financial products, such as funds, stocks and other financial products, suffered losses of varying degrees. Investors have also suffered significant property losses. The value investment and the technology trend investment also lose its function under the influence of the financial storm. At this time, the quantitative investment gradually shows its head and maintains a better return rate in the capital market. Indeed, with the continuous development of science and technology, artificial intelligence technology has become more mature, in order to quantify investors will be mathematical finance, applied statistics and other big data theory with computer, And the use of computers far superior to the human brain computing speed to find the entire financial market to achieve excess profit strategy laid the foundation. Today, with the increasing number of investment products, with the gradual improvement of China's securities market, the concept of investors is gradually transformed from the traditional technical analysis to the gradual vectorization of investment products, and quantitative investment has become an inevitable trend. In this paper, we use MATLAB software to obtain the fundamental and technical data of the stock from the WIND institutional edition, and eliminate some short listed companies, and take the whole A-share market stock as the sample to carry on the quantitative analysis. Finally, a reasonable quantitative investment strategy is given by taking the return rate as the standard of measurement. In this paper, we first introduce the related concepts of quantitative investment and the domestic and foreign development of quantitative trading, and review the literature; secondly, elaborate the process of multi-factor stock selection and its theoretical support; third, In this paper, the alpha multi-factor stock selection model is established by selecting the factors with good return rate in the back test. From the angles of valuation, growth, technical aspect and so on, 15 commonly used factors are selected as the factors to be selected. From January 2007 to December 2013, the validity and redundancy of these factors were tested, and three effective stock selection factors were obtained by comparing the return of portfolio with that of CS500 in the same time period. Combined with the statistical test method, a comprehensive multi-factor quantitative stock selection model is established, which lays a foundation for investors to find out a reasonable stock portfolio in line with the Chinese market. A quantitative trading platform based on MATLAB is built, and the stability and maneuverability of the platform are tested by using simple strategy to buy and sell stocks. When the platform is running, the historical data of the back test comes from the WIND institutional edition, the real-time trading data is obtained by the client of Sina stock market, and finally the stock API is attached to Tonghuashun to realize the full automatic trading of the stock. The main functions of the platform include: stock account login, buy, sell, query funds, query positions, query transactions, query entrustment, withdrawal and so on.
【学位授予单位】：天津商业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.51

【参考文献】