基于多分辨分析的沪深股市相关性分析

Analysis of SH&SZ Stock Market Dependence Based on Multiresolution Recognition

为了有效地揭示沪深股市不同交易周期股票交易的相关性,本文采用极大重叠离散小波变换,将上证指数和深圳成指高频收益率分解在不同的交易周期上,对各个周期的收益率采用semi-GPD模型作为其边缘分布分别进行拟合,在此基础上采用Copula函数方法建立同周期收益率的联合分布,度量了沪深两市同周期交易的相关性.实证表明,不同交易周期所表现出的相关性存在明显差异,并且随着交易期的增长,沪深两市非对称结构逐步明显。

To efficiently measure the correlation between different trading periods of SH and SZ stock market, based on decomposing the high frequency returns of two stock markets into different trading periods by MODWT, Copulas are studied by modifying the joint distributions of the same periods whose margins distributions are fitted by semi-GPD model. Evidences shows that obvious distinctness between each trading periods is in existence and the asymmetry structure of two markets is more escalate along with the increasing of trading periods.