相依结构数据的ACE拟合

Fitting the Data with Depedent Structure by Using ACE Method

利用ACE(alternating conditional expectations)方法,给出了寻找数据阿基米德copula相依结构的一种非参数算法.这种算法去掉了传统方法中对生成元需要事先假定结构类型的要求。拓宽了阿基米德copula的应用范围.模拟结果表明,ACE方法在寻找资产相依结构中是可以应用的,与传统方法相比也毫不逊色。

ACE method, as a nonparametric algorithm, is used to investigate the structure of Archimedean copulas. The method deletes the assumption of known structure which is needed for generator of Archimedean copulas in traditional methods, and extends the applied range of copulas. Simulation results show that ACE method is not worse than the traditional methods, and that ACE method is useful for identifying the dependent structure.