股票的交易数据拟合与聚类研究

发布时间：2018-02-02 07:31

本文关键词： 股票函数性数据数据拟合聚类　出处：《哈尔滨工业大学》2012年硕士论文　论文类型：学位论文

【摘要】：随着中国股票市场的迅速发展，市场的规范化程度不断提高，股票品种也有了向多元多层次化发展的趋势，吸引了越来越多投资者的目光。为减少投资风险，获得丰厚的利润回报，理智的股票投资者将会更加重视投资对象的选择。表达股票数据的真实意义对投资者来说是关键，而股票交易数据包含了大量的信息，对股票交易数据的分析就显得特别重要。股票交易数据的表现受很多因素的影响，包含信息量较大，总体上体现出函数性特征，采用传统的时间序列数据分析方法受到很多局限。为此，根据股票交易数据的函数性特征，借助函数性数据分析方法，对股票交易数据进行了有针对性的分析。主要内容是基于股票交易数据的函数性特征，对股票交易数据进行预处理和曲线拟合，，使得原始数据“抽象化”，进而得到统一的函数系数矩阵，再借助系数矩阵对反映股票特性的函数进行聚类，得出相应的股票聚类结果，并对结果进行了合理解释。将函数性数据分析方法应用于分析研究函数性数据中，改善了传统分析方法对数据要求的约束，这样不仅增加了可分析数据的范围，而且扩大了函数性数据分析方法的应用领域。将该方法实际应用于现实股票交易数据的拟合和聚类分析中，得到了非常理想的结果，表明了方法的有效性，该方法也能够为投资者提供很好的决策依据。
[Abstract]:With the rapid development of China's stock market, the standardization of the market is constantly improving, and the stock varieties have the trend of multi-level development, attracting the attention of more and more investors in order to reduce the investment risk. In order to get a good profit return, rational stock investors will pay more attention to the choice of investment object. Expressing the true meaning of stock data is key to investors, and stock trading data contains a lot of information. The analysis of stock trading data is particularly important. The performance of stock trading data is affected by many factors, including a large amount of information, which generally reflects the functional characteristics, and the traditional time series data analysis method is limited. According to the functional characteristics of the stock trading data and with the help of the functional data analysis method, the paper makes a targeted analysis of the stock trading data. The main content is based on the functional characteristics of the stock trading data. The preprocessing and curve fitting of the stock trading data make the original data "abstract", and then get the unified function coefficient matrix, and then cluster the functions reflecting the stock characteristics with the help of the coefficient matrix. The corresponding stock clustering results are obtained, and the results are explained reasonably. The functional data analysis method is applied to analyze and study the functional data, which improves the constraint of the traditional analysis method on the data requirement, which not only increases the scope of analytical data. Moreover, the application field of the functional data analysis method is expanded. The method is applied to the real stock trading data fitting and clustering analysis, and the very ideal result is obtained, which shows the validity of the method. This method can also provide investors with a good basis for decision-making.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F832.51;F224

【参考文献】