非线性因子分析在股票收益率模型中的应用

发布时间：2018-05-06 18:49

本文选题：因子分析 + 多因子模型　；参考：《华东师范大学》2017年硕士论文

【摘要】：本文运用量化投资中的多因子模型理论架构和方法,通过非线性的因子分析方法,将提取出的因子用于选股投资,力求赚取选股收益。通过线性和非线性两种因子分析结果和股票分层收益率的实证分析,比较两种分析方法的效果。全文一共分五个部分进行阐述。第一部分,简要介绍量化投资的基本概念和常用的量化选股模型,为本文的重点—多因子模型做铺垫。第二部分,详细介绍多因子模型的理论架构和方法,给出因子模型及其风险结构的数学表达。第三部分,回顾传统线性因子分析理论,推导因子载荷的求法,介绍因子旋转的方法和因子得分的求法。第四部分,引入非线性因子分析理论,着重介绍两种因子分析在思路和逻辑上的差别,通过证明正交因子定理以及引入贡献率的概念,完善因子分析的体系,使之可以应用到实际问题中。第五部分,对股票指标进行实证分析。首先对指标进行处理,然后分别用线性和非线性因子分析提取公共因子,根据这些因子对股票进行聚类,找出预期收益率最高的一类进行投资。最后对2014年至2015年中国沪深300成分股进行回测验证,比较两种分析方法的效果。
[Abstract]:In this paper, we use the theory and method of multi-factor model in quantitative investment, through nonlinear factor analysis method, to extract the factors for stock selection investment, and strive to earn stock selection income. The results of linear and nonlinear factor analysis and the empirical analysis of stock returns are compared. The full text is divided into five parts. In the first part, the basic concept of quantitative investment and the commonly used quantitative stock selection model are briefly introduced, which pave the way for the multi-factor model of this paper. In the second part, the theoretical framework and method of the multi-factor model are introduced in detail, and the mathematical expression of the factor model and its risk structure is given. In the third part, the traditional theory of linear factor analysis is reviewed, the calculation method of factor load is deduced, the method of factor rotation and the method of factor score are introduced. In the fourth part, the theory of nonlinear factor analysis is introduced, and the differences between the two kinds of factor analysis in thought and logic are introduced. By proving the orthogonal factor theorem and introducing the concept of contribution rate, the system of factor analysis is improved. So that it can be applied to practical problems. The fifth part, carries on the empirical analysis to the stock index. First, the indexes are processed, then the common factors are extracted by linear and nonlinear factor analysis, according to these factors, the stocks are clustered to find out the highest expected return rate of investment. Finally, the back test of China's Shanghai and Shenzhen 300 component stock from 2014 to 2015 is carried out to compare the results of the two analysis methods.
【学位授予单位】：华东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F832.51;F224

【参考文献】