CAJ | 학술논문

本文以沪深300股指期货与上证指数的日收益率序列作为建模对象,通过两种基于经验分布的半参数估计copula参数的方法对两序列进行了相关性分析.结果表明,通过QQ、K-S拟合检验以及欧式距离,这两种方法所估计的Gumbel-copula模型均能更好地描述沪深300股指期货与上证指数的相关关系.
본문이호심300고지기화여상증지수적일수익솔서렬작위건모대상,통과량충기우경험분포적반삼수고계copula삼수적방법대량서렬진행료상관성분석.결과표명,통과QQ、K-S의합검험이급구식거리,저량충방법소고계적Gumbel-copula모형균능경호지묘술호심300고지기화여상증지수적상관관계.
Copula theory plays an important role in solving complex problems in the areas of financing and economy. It is important to estimate the parameter of copula function,when using it in modeling of multi-variate series. Semiparametric estimation methods are not affected by marginal distribution and robust,thus they are popularly used. Yu Bo proposes a new semiparametric estimation method based on BFGS algorithm. Using semiparametric estimation of Copula parameter based on empirical distribution,this paper implements correlation analysis on two daily yield series（ CSI 300 and SSE）. The analysis indicates that Gumbel-copula is the best model for correlation between CSI 300 and SSE.