黄金期货市场非线性特征及相关问题研究

发布时间：2018-02-08 19:04

本文关键词： 黄金期货分形分析赫斯特指数多重分形 FIAPARCH模型　出处：《浙江工商大学》2012年硕士论文　论文类型：学位论文

【摘要】：黄金期货市场是一个复杂的系统,同时也是一个非线性系统。有效市场假说(EMH)以线性化的假设来刻画市场,未能准确反映市场的真实特性,而分形市场假说(FMH)的提出可以解决这一问题。本文依据分形市场假说的思想,对上海期货交易所和纽约商品交易所的黄金期货收益率数据作了实证研究,从统计学的角度,定量地证明上述两黄金期货市场实际上具有分形结构特征,并在分形理论的基础上解释黄金期货市场的特性。此外,为了了解黄金期货在不同市场及不同时期在分形特征上的异同,本文还对数据进行分段对比研究以求达到更完善的研究目的。本文首先对上海、纽约黄金期货收益率序列的基本统计特征及非线性特征进行检验,并在此前提下进行分形特征研究(包括单分形特征及多重分形特征研究)。在单分形特征分析中,主要运用重标极差分析方法(R/S)计算Hurst指数来分析黄金期货收益率及其波动序列的分形和长记忆性特征。之后,针对上海、纽约黄金期货收益率波动序列具有长记忆性特征,本文引入正态分布、t分布、偏t分布下的带长记忆性的FIGARCH模型和FIAPARCH模型对上海、纽约黄金期货收益率序列进行计量分析,以此为两市黄金期货选择最恰当的波动模型并作相关波动特征的解释。而为了弥补单分形分析的不足,本文进一步对黄金期货进行多重分形特征分析,通过运用消除趋势波动分析方法(MF-DFA)计算随着阶数q值变化的广义Hurst指数来更加细致刻画黄金期货市场的持续性等特征变化。通过本文的研究表明运用分形市场假说及相关分形分析方法对黄金期货进行研究是有成效的,这让我们更好地了解黄金期货市场的内在特性及其演变特征,也为我国黄金期货市场的发展提供更多的理论依据。
[Abstract]:The gold futures market is a complex system and a nonlinear system. The efficient Market hypothesis (EMH) characterizes the market by linearization and fails to accurately reflect the true characteristics of the market. The fractal market hypothesis (FMH) can solve this problem. According to the idea of fractal market hypothesis, this paper makes an empirical study on the data of gold futures yield of Shanghai Futures Exchange and New York Mercantile Exchange. It is proved quantitatively that the above two gold futures markets actually have fractal structural characteristics, and the characteristics of gold futures markets are explained on the basis of fractal theory. In order to understand the differences and similarities of gold futures in different markets and in different periods, this paper also makes a comparative study of the data in order to achieve a more perfect research purpose. This paper first examines the basic statistical characteristics and nonlinear characteristics of gold futures yield series in Shanghai and New York. In this context, fractal features (including single-fractal features and multifractal features) are studied. This paper mainly uses the method of rescaled range analysis to calculate the Hurst index to analyze the fractal and long memory characteristics of gold futures yield and its volatility series. Then, the volatility series of gold futures yield in Shanghai and New York have long memory characteristics. In this paper, we introduce the FIGARCH model with long memory and the FIAPARCH model under normal distribution and partial t distribution to analyze the gold futures yield series in Shanghai and New York. In order to make up for the deficiency of single fractal analysis, this paper further analyzes the multifractal characteristics of gold futures. The generalized Hurst exponent with order Q change is calculated by using trend elimination analysis method (MF-DFAA) to characterize the persistence of gold futures market in more detail. Through the research in this paper, it is shown that it is effective to use fractal market hypothesis and relevant fractal analysis methods to study gold futures, which makes us better understand the inherent characteristics and evolution characteristics of gold futures market. It also provides more theoretical basis for the development of China's gold futures market.
【学位授予单位】：浙江工商大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F831.54;F713.35;F224

【参考文献】