股票定价函数形式及其数据去噪方法的研究

发布时间：2018-06-03 22:15

本文选题：定价函数 + 变量　；参考：《中国计量学院》2013年硕士论文

【摘要】：价格作为买卖双方最重要的信号，是市场的灵魂，因此构建合理的股票定价函数显得尤为重要。定价函数形式的研究也日益受到众多学者的青睐，提出了各种定价方法。但是，目前的研究成果仍有许多待改进的地方，如选取变量单一、函数形式多为线性等。结合股票价格存在噪声等问题，本文在已有文献基础上，主要研究了以下三方面的内容：（1）本文通过股票定价机制的分析，认为目前定价函数所使用的每股收益、每股净资产等公司层面的变量有改进的空间，提出公司层面应该与市场层面相结合，通过分析将Beta值与流通股本选为市场层面变量，并进行了实证分析。结果表明，综合考虑公司与市场层面变量比仅考虑公司层面变量在模型的预测精度上更有优势。（2）通过多种形式的生产函数对比分析，本文认为由于股票定价机制的复杂性，股票定价函数形式应该是非线性的。综合比较，，本文将柯布-道格拉斯生产函数选为定价函数。同时，介绍了对回归模型的统计检验方法，为后面的实证分析做理论依据。在实证部分，本文将线性形式的定价函数与非线性形式的定价函数进行了对比研究，结果显示，非线性模型的预测精度明显好于线性模型的预测精度。（3）股票价格作为金融时间序列的代表，其噪声问题不可回避，因此去噪问题的研究十分重要。在此基础之上，本文提出了基于小波分析的去噪方法，并进行实证研究，结果表明，经过去噪之后的模型预测精度明显得到提升。
[Abstract]:Price, as the most important signal between buyer and seller, is the soul of market, so it is very important to construct reasonable stock pricing function. The research of pricing function form has been increasingly favored by many scholars, and various pricing methods have been put forward. However, there are still many improvements in the present research results, such as single variables and linear functions. Based on the existing literature, this paper mainly studies the following three aspects: 1) through the analysis of stock pricing mechanism, this paper thinks that there is room for improvement in the variables of company level such as earnings per share, net assets per share and so on, which are used in the current pricing function, and puts forward that the company level should be combined with the market level. Through the analysis, the Beta value and the circulating stock are selected as the market level variables, and the empirical analysis is carried out. The results show that the prediction accuracy of the model is better when considering the variables at the firm and market level comprehensively than only the variables at the firm level. 2) based on the comparative analysis of various production functions, this paper holds that due to the complexity of stock pricing mechanism, the form of stock pricing function should be nonlinear. In this paper, Cobb-Douglas production function is selected as the pricing function. At the same time, the statistical test method of regression model is introduced, which is the theoretical basis for the later empirical analysis. In the empirical part, we compare the linear pricing function with the nonlinear pricing function. The results show that the prediction accuracy of the nonlinear model is obviously better than that of the linear model. As the representative of financial time series, the noise problem of stock price can not be avoided, so it is very important to study the problem of de-noising. On this basis, this paper proposes a method of denoising based on wavelet analysis, and carries out an empirical study. The results show that the prediction accuracy of the model is obviously improved after denoising.
【学位授予单位】：中国计量学院
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F830.91

【参考文献】