基于回归法的多因子选股模型的投资组合分析

发布时间：2018-05-08 14:37

本文选题：回归法 + 多因子选股模型　；参考：《青岛大学》2017年硕士论文

【摘要】：随着量化投资概念在世界各地兴起,量化投资也越来越多的受到中国证券业界的关注。与此同时,量化投资的思想和理念已经逐渐被市场所接受,开始成为证券业界研究的热点问题。虽然量化投资理念在中国刚刚起步,但是在中国市场中展现了较强的生命力和发展空间。在中国金融改革不断推进的背景下,对量化投资的深入研究也是证券业发展的大势所趋。本文介绍的多因子模型是目前国际上主流的量化投资模型之一,也是目前中国量化投资领域的热点问题。多因子模型试图通过建模的方法对驱动股票市场价格变化的因子进行解释和分析。多因子模型的研究也将对券商和投资基金的运作具有一定的指导意义。目前国内流行的很多量化投资模型都是建立多因子模型的框架基础上,因此研究多因子模型的有效建模方法是目前量化投资领域业界关注的一个重要问题。本文基于MATLAB和EXCEL软件和锐思数据库利用回归法构建了多因子选股模型。本文选取的样本数据为2011-2016年间具有A股代表性的沪深300股票市场上的财务指标和其他因子指标。对于因子的选取则考虑了市场的经验以及因子对公司的代表性。本文首先对因子进行了初步的单因子检验,然后通过Fama-Mac Bech检验删除了表现不显著的部分因子;接下来计算了剩余因子的相关系数矩阵,剔除了具有多重共线性的同类型因子;然后利用MATLAB进行逐步回归,得到了最终的回归选股模型。在实证检验环节创新性的加入了约束条件对股票组合的个股进行了赋权,并利用滚动检验的方式对持有期的四组股票组合的收益进行了检验,证实了模型选股的有效性和实际参考价值。本文主要偏重于多因子模型的实践与应用,对目前理解和正确运用多因子模型具有十分重要的意义。希望本文能够让读者对多因子选股有更深入的了解和认识。
[Abstract]:With the rise of the concept of quantitative investment in various parts of the world, quantitative investment has been paid more and more attention by Chinese securities industry. At the same time, the thought and idea of quantitative investment has been accepted by the market gradually, and has become a hot issue in the securities industry. Although the concept of quantitative investment in China has just started, but in the Chinese market has shown a strong vitality and development space. Under the background of China's financial reform, the deep research on quantitative investment is also the trend of the development of securities industry. The multi-factor model introduced in this paper is one of the mainstream quantitative investment models in the world at present, and is also a hot issue in the field of quantitative investment in China. The multi-factor model attempts to explain and analyze the factors driving the change of stock market price by modeling. The study of multi-factor model will also have certain guiding significance to the operation of securities firms and investment funds. At present, many popular quantitative investment models are based on the framework of multi-factor model, so it is an important problem to study the effective modeling method of multi-factor model in the field of quantitative investment. In this paper, a multi-factor stock selection model is constructed based on MATLAB, EXCEL software and Rilith database. The sample data selected in this paper are the financial indexes and other factor indexes in the Shanghai and Shenzhen 300 stock markets, which are representative of A shares in 2011-2016. For the selection of factors, we consider the market experience and the representation of the factors to the company. In this paper, we first test the factors by single factor, then delete the unremarkable factors by Fama-Mac Bech test, then calculate the correlation coefficient matrix of the remaining factors, and eliminate the same type factors with multiple collinearity. Then using MATLAB stepwise regression, the final regression stock selection model is obtained. In the link of empirical testing, we add the constraint conditions to empower the stock portfolio, and use the rolling test to test the earnings of the four groups of stock portfolio in the holding period. The validity and practical reference value of model stock selection are confirmed. This paper focuses on the practice and application of the multi-factor model, which is of great significance to understand and correctly use the multi-factor model at present. I hope this article will enable readers to have a deeper understanding and understanding of multi-factor stock selection.
【学位授予单位】：青岛大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.51

【参考文献】