沪铜期货市场分形特征研究

发布时间：2018-05-06 19:34

本文选题：沪铜期货 + 分形　；参考：《北方工业大学》2014年硕士论文

【摘要】：应用自相关函数检验法和GARCH模型检验方法,对时间标度τ=1,5,22,66下的沪铜期货收益率序列进行实证研究,在不同时间标度下均检验出波动聚集现象,沪铜期货表现出自相似的特征。通过移动窗口技术构建波动聚集指数,度量了沪铜期货波动聚集程度,发现其波动聚集程度随着T和移动窗口的增大而增大。利用重标极差(R/S)方法和修正的重标极差方法(修正R/S分析法)对沪铜期货市场不同时间标度下的收益率序列进行实证分析,并将分析结果进行了对比。研究发现修正R/S分析法在不同时间标度下没有检验出沪铜期货市场具有明显的长期记忆的特征,而经典的R/S分析法支持沪铜期货市场具有持续性和记忆性的特征。运用去趋势波动分析法(DFA),对沪铜期货不同时间标度下的收益率序列的长记忆特征进行研究,并对沪铜期货市场的标度特性进行了分析。结果表明,沪铜期货收益率序列的标度指数和长期记忆性随着时间标度的增大而增大。趋势阶数较高时,不同时间标度下的收益率序列的标度指数近似相等。运用多重分形去趋势波动分析法(MF-DFA)对沪铜期货市场不同时间标度收益率序列进行实证研究,发现其具有显著的多重分形特征。通过对原始序列作重排和相位随机化处理,并与原始收益率序列进行比较,深入分析了多重分形特征的成因。结果表明,沪铜期货收益率序列的多重分形特征是由长期记忆性和肥尾分布两个因素导致的,其中价格波动的长期记忆性是主要原因。
[Abstract]:Using autocorrelation function test method and GARCH model test method, this paper makes an empirical study on the Shanghai copper futures yield series under time scale 蟿 ~ (1) ~ (1) ~ (5) ~ (5) ~ (2) ~ (2 +). Under different time scales, the phenomenon of fluctuation aggregation is tested, and the performance of Shanghai copper futures is similar. Based on the moving window technology, the volatility aggregation index is constructed to measure the volatility aggregation degree of Shanghai copper futures. It is found that the volatility aggregation degree increases with the increase of T and moving window. By using the method of rescaling range difference and the modified method of rescaling range (modified R / S method), the paper makes an empirical analysis on the return series of Shanghai copper futures market under different time scales, and compares the results. It is found that the modified R / S method does not test the characteristics of long term memory in Shanghai copper futures market under different time scales, while the classical R / S analysis supports the persistence and memory of Shanghai copper futures market. The long memory characteristics of the yield series of Shanghai copper futures under different time scales are studied by using the detrend volatility analysis method (DFAA), and the scale characteristics of Shanghai copper futures market are analyzed. The results show that the scale index and long-term memory of Shanghai copper futures yield series increase with the increase of time scale. When the trend order is high, the scalar exponents of the return series under different time scales are approximately equal. The multifractal detrend volatility analysis method (MF-DFAA) is used to study the different time scale return series of Shanghai copper futures market, and it is found that it has significant multifractal characteristics. By rearranging and phase randomizing the original sequence and comparing it with the original rate of return series, the causes of multifractal characteristics are analyzed in depth. The results show that the multifractal characteristics of Shanghai copper futures return series are caused by two factors: long-term memory and fat tail distribution, among which the long-term memory of price fluctuation is the main reason.
【学位授予单位】：北方工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F724.5;F224

【参考文献】