基于股市交易量价分布的状态跃迁模型研究

发布时间：2018-05-13 21:18

本文选题：量价分布 + 状态跃迁　；参考：《暨南大学》2014年硕士论文

【摘要】：随着计算机处理能力的提高和大数据时代的到来，数学、物理以及数据挖掘的计算机方法正逐渐广泛应用于经济学和金融学的研究，在一定程度上解决了传统经济学和金融学定性研究和定量研究方法不足，以及不符合社会实际的理性人假设与有效市场假说的缺陷，给经济学和金融学点亮了新的研究视野和方向。本文结合数理方法和数据挖掘的技术，探讨了中国股市交易的量价分布的近似正态分布形态，研究了中国股市交易量价分布不同时间跨度的近似正态分布形态结构，并在此基础上进一步提出了中国股市交易量价分布的预测模型。本文的工作主要包括三个部分：首先，验证了中国股市日交易量价分布拒绝服从正态分布的假设，同时指出，中国股市日交易量价分布由于局部成交量较大，破坏了正态分布的形态，但是展现出“尖峰厚尾”的近似正态分布形态。然后，经过对量价分布的平滑处理，得到了股市交易量价分布的无标度微观形态——不同时间跨度的近似正态分布形态结构。最后，，在前两个研究的基础上，提出了股市量价分布的状态跃迁模型。使用聚类分析，建立基于概率的状态跃迁表，预测未来可能出现的近似正态分布形态微观结构。本文所建立的量价分布状态跃迁模型可以适当的描述中国股市交易中量价分布状态之间的跃迁网络，并在一定概率上预测未来可能出现的量价分布状态，为具体的股市投资提供一个具体量化的指标。
[Abstract]:With the improvement of computer processing ability and the arrival of big data era, computer methods of mathematics, physics and data mining are being widely used in the research of economics and finance. To some extent, the defects of the traditional qualitative and quantitative research methods in economics and finance are solved, and the rational man hypothesis and the efficient market hypothesis, which do not conform to the social reality, have been solved. To economics and finance to light up a new perspective and direction of research. Combined with mathematical methods and data mining techniques, this paper discusses the approximate normal distribution of volume and price distribution in Chinese stock market, and studies the approximate normal distribution structure of trading volume price distribution in China stock market with different time span. On the basis of this, a forecasting model of the price distribution of trading volume in Chinese stock market is put forward. The work of this paper mainly includes three parts: firstly, it verifies the hypothesis that the daily transaction price distribution of Chinese stock market is based on the normal distribution, and points out that the price distribution of daily trading volume in Chinese stock market is due to the large local volume. The shape of normal distribution is destroyed, but the shape of "peak thick tail" is shown. Then, by smoothing the volume price distribution, the scale-free microscopic morphology of the price distribution of the stock market is obtained, which is the approximate normal distribution structure with different time span. Finally, on the basis of the first two studies, the state transition model of stock price distribution is proposed. Using clustering analysis, a probabilistic state transition table is established to predict the morphology of approximate normal distribution in the future. The model can describe the transition network between the volume and price distribution states in the Chinese stock market, and predict the potential price distribution in the future in a certain probability. For the specific stock market investment to provide a specific quantitative indicators.
【学位授予单位】：暨南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F832.51

【参考文献】