基于特质波动率的金融时间序列挖掘建模研究

发布时间：2018-03-13 03:32

本文选题：时间序列　切入点：数据挖掘　出处：《暨南大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着计算机技术、人工智能、机器学习以及统计分析方法的有机融合和发展，数据挖掘技术得到了迅猛发展，且伴随着大数据时代的到来，传统的金融分析方法已经逐渐无法满足金融数据分析的应用和要求，采用数据挖掘的方法对金融时间序列数据进行分析逐渐成为了金融研究的潮流。在这一背景下，本文对金融时间序列中的股价时间序列数据进行挖掘建模研究，同时，鉴于目前特质波动率对股价趋势的预测这方面的研究还相对较少，因此，本文采用自定义的TBUD方法对股价时间序列进行集合划分，并采用支持向量机进行建模，研究集合的特质波动率属性对趋势的预测能力。本文的实证研究发现，，TBUD方法所划分的集合之间的特质波动率属性差异并不显著，特质波动率无法对股价趋势做出准确的预测。本文提出时间序列上的拐点集合划分方法，摆脱从回归方程上来对时间序列进行预测，而是从数据挖掘的角度，研究拐点集合与趋势的相关性，为以后的研究提供一个新的方向。
[Abstract]:With the integration and development of computer technology, artificial intelligence, machine learning and statistical analysis methods, data mining technology has been rapidly developed, and accompanied by the arrival of big data era. The traditional method of financial analysis has been unable to meet the application and requirement of financial data analysis. Using the method of data mining to analyze the financial time series data has gradually become the trend of financial research. In this context, this paper studies the mining and modeling of stock price time series data in financial time series. At the same time, in view of the fact that there is relatively little research on the prediction of stock price trend by idiosyncratic volatility, therefore, In this paper, we use the custom TBUD method to partition the stock price time series, and use support vector machine to model the stock price time series. In this paper, we find that there is no significant difference in the trait volatility attributes between the set and the TBUD method, and the trait volatility can not accurately predict the stock price trend. In this paper, a method of dividing the inflection point set in time series is proposed to predict the time series from the regression equation, but the correlation between the inflection point set and the trend is studied from the angle of data mining. It provides a new direction for future research.
【学位授予单位】：暨南大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F830.91

【参考文献】