我国沪铜期货价格的分形统计

Fractal Statistics for Copper Price on Shanghai Futures Market

运用分形分布参数估计、函数盒维数以及多重分形分析等方法对我国沪铜期货价格时间序列进行了实证研究.结果表明,沪铜期货价格不服从正态分布,价格之间存在长记忆性,从而对有效市场假说提出了质疑.函数盒维数及多标度分析的结果揭示了期货价格的聚类特征及标度变化,说明用单一分形模型来描述期货价格是不充分的,多重分形分析方法为更好地描述期货价格的变化规律提供了有力的工具.

Parameters＇ estimation of fractal distribution, functional box dimension and multifractal analysis are used to study empirically the fractal characteristics of the time series of copper price on Shanghai futures market. It is found that the time series are not normally distributed and show long-term memory. The efficient market hypothesis is thus to be queried. The clustering structure and scale variation of the time series as revealed by the functional box dimension and multifractal analysis show that a single fractal model is not sufficient to describe the futures price in time series. However, the multifractal analysis as a powerful instrument can describe more exactly how the price of copper futures varies.