关键词: C-Vine Copula, 广义自回归得分, 时变参数, 动态相关结构

Abstract: It is a hot topic for both academic and practical fields to describe the nonlinear dynamic dependence structure among high dimensional financial variables accurately. In this paper, based on Generalized Auto-regressive Score (GAS) theory, we propose a new semi-parametric method in estimating the time-varying parameters and a goodness of fit test of the time-varying C-Vine Copula model. It shows that the model could capture the nonlinear and dynamic dependence structure among random variables. The reliable and robust performances of the proposed method are further illustrated by Monte Carlo simulation.

Key words: C-Vine Copula, Generalized Auto-regressive Score, Time-Varying Parameter, Dynamic Dependence Structure