主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2008, Vol. 20 ›› Issue (6): 24-28.

• 论文 • 上一篇    下一篇

金融ARCH模型的贝叶斯检验和模型选择

李勇1, 倪中新2   

  1. 1. 中山大学管理学院财务与投资系, 广州 510275;
    2. 上海大学国际工商与管理学院金融系, 上海 200444
  • 收稿日期:2008-01-23 修回日期:2008-10-13 出版日期:2008-12-31 发布日期:2008-08-20
  • 作者简介:李勇(1979- ),男(汉族),江苏南京人,中山大学管理学院讲师,香港中文大学博士,研究方向:贝叶斯金融计量经济学、金融工程.

Bayesian Testing and Model Comparison for Financial ARCH Models

LI Yong1, NI Zhong-xin 2   

  1. 1. Business School, Sun Yat-Sen University, Guangzhou 510275, China;
    2. College of International Business and Management, Shanghai University, Shanghai 200444, China
  • Received:2008-01-23 Revised:2008-10-13 Online:2008-12-31 Published:2008-08-20

摘要: 在金融时间序列分析中,检验ARCH效果和决定合适的阶是ARCH模型的重要研究主题,在贝叶斯框架下,本文使用贝叶斯因子来检验ARCH效果和选择ARCH模型合适的阶。在路径抽样的基础上,提出了计算ARCH模型贝叶斯因子的方法。最后,我们用一个具体的实例来论证了所提方法的有效性。

关键词: ARCH模型, 贝叶斯因子, 金融时间序列, ARCH效果检验, 模型选择, 路径抽样

Abstract: In financial time series analysis,testing the ARCH effect and determining the appropriate order value is one of the important topics. In this paper,the Bayes factor is employed to test the ARCH effect and choose the appropriate order value for ARCH models under Bayesian framework. A procedure for computing Bayes factor based on path sampling is established for this purpose. In the end,a real example is illustrated to demonstrate our proposed method.

Key words: ARCH models, ARCH effect, Bayes factor, financial time series, path sampling, model comparison

中图分类号: