动态Copula模型允许变量间的相关性随时间或其它协变量的变化而变化，因此能够描述相关性结构的变化趋势。动态混合Copula模型是若干动态Copula模型的加权平均，由于其权重参数也可以随时间或其它协变量的变化而变化，因而大大增强了模型的准确性和灵活性。本项目以动态混合Copula模型为研究对象，建立起一套基于数据驱动的模型估计和选择方法。本项目具体内容包括：构造非参数动态混合Copula模型并对模型中的未知参数加以估计，并于此同时对组成混合Copula模型的各个Copula加以选择；得到估计值的渐近性质；利用蒙特卡洛模拟对提出的估计和选择方法在有限样本下的表现进行验证；运用构造的模型和方法进行实证分析，如分析两岸金融市场间的相关性结构受两岸贸易等变量的影响而呈现的变化趋势。本项目旨在提供探究金融市场相关性结构的模型和估计选择方法，为进一步提高金融风险管理和预测的准确度奠定理论和技术基础。

Time-varying copula models allow the dependence structure among variables to change with time or covariates, and can describe the evolution of the dependence structure. A time-varying mixture copula is formulated as a weighted average of several individual time-varying copulas. Because the weight parameters can change with time or covariates as well, a time-varying mixture copula offers additional flexibility and accuracy. This research project focuses on time-varying mixture copula models and proposes a data-driven copula estimation and selection method. Specifically, in the project, we propose a nonparametric time-varying mixture copula model, estimate the parameters and select the copula components simultaneously, establish the asymptotic properties of the proposed estimator, and illustrate the finite sample performance of the proposed method by Monte Carlo simulation. Finally, we apply the proposed approach to empirical studies. For example, we investigate the influence of the cross-Strait trade on the dependence between the financial markets of Taiwan and Mainland China. This research project aims to provide a model and an estimation and selection method to study the dependence among financial markets, and lay a theoretical and technical foundation for the further development of financial risk management and prediction.