中文摘要：

为了分析我国与国际上其他已推出股指期货的国家或地区股指波动特征的相似性，采取传统时间序列模型分析与数据挖掘技术相结合的方法，对全球23个国家或地区的股指波动特征作了聚类分析。首先，针对股指收益序列的非对称性和异方差特性，建立非对称效应异方差模型并估计其模型系数。然后，对特征抽取后的系数使用欧氏距离判断序列之间的相似程度并进行层次聚类分析。最后，通过实证检验各个国家或地区的股指波动相似性，找出了与我国情况较为近似的国家或地区，从而证实了本文方法的有效性。

英文摘要：

In order to analyze the similarity of volatility of stock indices between China and other counties or areas that have already introduced stock index futures, this paper utilizes traditional time series model analysis and data mining method to perform clustering analysis on stock index volatilities of 23 countries or areas. Firstly, asymmetrical heteroskedastic models are developed for the asymmetric and heteroskedastic properties of stock return time series, and the coefficients are then estimated. Secondly, Euclidean dis- tance is used to justify the degree of similarity of the features extracted from the coefficients series, and hierarchical clustering is performed also. Finally, by empirically examining the similarity of stock index volatilities in each country or area, those countries or areas that are similar to that of China are found. The effectiveness of this method is thus demonstrated.