基于符号时间序列分析的多尺度金融波动研究

发布时间：2018-04-05 20:24

本文选题：多尺度分析　切入点：符号时间序列分析　出处：《天津大学》2012年硕士论文

【摘要】：金融波动性是金融市场的内在属性，与金融市场的功能与稳定性密切相关，也是衡量一个市场效率和发展完善程度的指标。金融市场的波动性意味着市场中不确定性和风险，因此无论对于市场投资者还是市场监管者而言，研究金融波动特性，全面、正确认识金融波动，把握金融波动特性，都有着重要的意义。金融市场本质上是一个非线性系统，影响因素众多，变化复杂。分形、混沌分析和符号时间序列分析等非线性系统分析方法逐步引入金融市场的分析中，以期能够更有效、更全面地揭示金融波动性的规律。众多研究结果发现，金融市场波动存在多尺度现象，利用单一时间尺度分析金融波动得到的结果往往是片面的。因此本文从多尺度的角度出发，采用符号时间序列分析方法，对金融市场的波动特性展开研究，希望能够全面且准确地认识金融波动的特性。首先，文章论述了进行金融市场波动研究的背景及意义，总结了金融波动的研究成果，重点介绍了符号时间序列分析方法和小波多分辨分析的基本理论，为论文的展开提供理论基础。然后将小波多分辨分析与符号时间序列方法结合，对将上证综指和深证成指的“已实现”波动序列分解为不同尺度的细节分量，对原序列及不同的细节分量分别采用符号化分析，得出不同尺度上的符号序列直方图，辨别确定不同时间尺度上的主要变化模式与异常变化模式，，为不同类型的投资者提供投资策略和风险管理的依据；提出符号序列秩次图，直观地研究不同序列之间以及同一序列在不同尺度上的相似性与差异性，体现符号时间序列分析的优越性，计算简便，减少了很多不必要的麻烦；采用欧几里得范数、2统计量、相对熵以及秩次距离等符号时间序列的统计量，定量地分析不同序列之间不同尺度上的差异性，用具体的值描述差异性。本文是国家自然科学基金项目“基于符号时间序列分析的金融波动研究”(项目编号：70971097)研究工作的一部分。
[Abstract]:Financial volatility is the intrinsic attribute of financial market, which is closely related to the function and stability of financial market, and is also an index to measure the market efficiency and the degree of development and perfection.The volatility of the financial market means the uncertainty and the risk in the market. Therefore, no matter for the market investors or the market regulators, we should study the characteristics of the financial volatility, understand the financial volatility correctly, and grasp the characteristics of the financial volatility.Are of great significance.Financial market is essentially a nonlinear system, which has many factors and complex changes.The nonlinear system analysis methods such as fractal, chaotic analysis and symbolic time series analysis are introduced into the analysis of financial market step by step, in order to reveal the law of financial volatility more effectively and comprehensively.Many studies have found that there are multi-scale phenomena in financial market volatility, and the results obtained by single time scale analysis are often one-sided.Therefore, from the point of view of multi-scale, this paper uses the symbolic time series analysis method to study the volatility characteristics of financial markets, hoping to fully and accurately understand the characteristics of financial volatility.First of all, the paper discusses the background and significance of financial market volatility research, summarizes the research results of financial volatility, focuses on the symbolic time series analysis method and wavelet multi-resolution analysis of the basic theory.To provide the theoretical basis for the development of the paper.Then the wavelet multi-resolution analysis is combined with the symbolic time series method to decompose the "realized" wave series of the Shanghai Composite Index and the Shenzhen Composite Index into the detail components of different scales.The symbol histogram of the original sequence and the different detail components is obtained by symbolic analysis on different scales, and the main change patterns and anomalous variation patterns on different time scales are identified by distinguishing the histogram of the original sequence and the different detail components, and the histogram of the symbol sequence on different scales is obtained.It provides the basis of investment strategy and risk management for different types of investors, and puts forward the rank graph of symbol sequence to study the similarity and difference between different sequences and the same sequence on different scales.The advantages of symbolic time series analysis are reflected, the calculation is simple and many unnecessary troubles are reduced, and the statistics of symbol time series, such as Euclidean norm, relative entropy and rank distance, are used.The differences between different scales are analyzed quantitatively and the differences are described with specific values.This paper is a part of the research work of the National Natural Science Foundation of China "Research on Financial volatility based on symbolic time Series Analysis" (item No.: 70971097).
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F830.91

【引证文献】