摘要
本文从Spearmanρ入手,利用Spearmanρ在非线性单调变换的情况下保持不变的特点,以及与条件期望预测机制存在的非线性的关系,提出建立时变Copula的模型的新方法;通过建立时变FGM-Copula模型的实例分析表明,这种构建Copula模型的方法较好捕捉了相依机制的时变性,预测了随机变量的趋势,具有一定的优越性。
In this paper, based on the Spearman p which remains invariability in non-linear monotonic transformation and has non-linear relationship with the prediction mechanism of conditional expectation, the new method of establishing the time-varying Copula model is proposed. Through the example of establishing the time-varying FGM-Copula model and the analysis, it is shown that this method of establishing a time-varying Copula model can capture the time-varying dependent mechanism, and predicted the trend of random variables. The method has certain superiority.
出处
《数理统计与管理》
CSSCI
北大核心
2011年第1期76-84,共9页
Journal of Applied Statistics and Management
基金
2009教育部人文社会科学研究项目基金资助(编号09YJCZH104)
中央高校基本科研业务费专项资金资助(编号SWJTU09CX075
SWJTU09ZT37)
关键词
时变COPULA
条件期望预测机制
Spearman
ρ回归函数
time-varying copula, the prediction mechanism of conditional expectation, spearman ρ, Copula regression functions