基于动态因子模型的金融风险指数构建

发布时间：2018-02-24 18:11

本文关键词： 金融风险指数动态因子模型正向客观赋权法　出处：《统计与决策》2017年20期 　论文类型：期刊论文

【摘要】：文章重点研究金融风险指数的构建方法,选取宏观维度、银行与货币维度、泡沫维度、外部冲击维度、债务维度等17个指标,运用动态因子模型方法构建2004—2015年的金融风险指数,并利用局部加权回归思想,引入高斯核函数改进指标权重估计,与常用的正向客观赋权法SF赋权法、熵赋权法、CRITIC赋权法对比。研究结果表明:改进后的权重估计方法拟合精度提高50%以上,动态因子模型方法构建的金融风险指数比正向客观赋权法灵敏度高,更科学更合理。
[Abstract]:This paper focuses on the construction method of financial risk index, selects 17 indexes, such as macro dimension, bank and currency dimension, bubble dimension, external shock dimension, debt dimension and so on, and uses dynamic factor model to construct financial risk index from 2004 to 2015. By using the idea of local weighted regression, Gao Si kernel function is introduced to improve the index weight estimation, which is compared with the SF weighting method, which is usually used as a positive objective weighting method. The research results show that the fitting accuracy of the improved weight estimation method is improved by more than 50%, and the financial risk index constructed by the dynamic factor model method is more sensitive than that of the positive objective weighting method and is more scientific and reasonable.
【作者单位】：武汉理工大学理学院;
【基金】：教育部人文社会科学基金资助项目(12YJAZH022)
【分类号】：F224;F832

【相似文献】