R-vine copula模型与PCBN模型的比较

发布时间：2018-04-04 22:31

本文选题：R-vine　切入点：copula　出处：《统计与决策》2016年23期

【摘要】：文章对比了两类刻画高维变量相依结构模型——R-vine copula模型和PCBN模型,并将其应用于国民经济九大行业信用风险相依结构分析,结果表明,与R-vine copula模型相比,PCBN模型能更好地兼顾模型的准确性和简洁性目标。通过PCBN模型可以发现:国民经济整个系统内行业间存在条件独立关系,其中七个行业构成的子系统是整个系统内风险传染的关键媒介,而在子系统内部,水电燃气、批发零售、信息软件及金融业是信用风险传染的关键媒介。
[Abstract]:This paper compares two kinds of models, R-vine copula model and PCBN model, which describe the dependence structure of high-dimensional variables, and applies them to the analysis of dependence structure of credit risk in nine industries of national economy. The results show that,Compared with the R-vine copula model, it can better balance the accuracy and conciseness of the model.Through the PCBN model, we can find that there are conditional independent relationships among industries in the whole national economy system, in which seven industries are the key media of risk contagion in the whole system, and within the subsystem, water and electricity gas, wholesale and retail,Information software and financial industry are the key media of credit risk contagion.
【作者单位】：南京工业大学数理科学学院;南京航空航天大学经济管理学院;
【基金】：国家自然科学基金资助项目(71401074) 江苏省哲学社会科学基金重点项目(14GLA003) 江苏省高校研究生科研创新计划项目(KYZZ_0099) 江苏省教育厅高校哲学社会科学研究项目(2016SJB630030)
【分类号】：F224

【相似文献】