基于支持向量机的银行系统重要性评估研究
Research on the Systemically Important Evaluation of Banks Based on Support Vector Machine
系统重要性银行是构成全球业务链的连接点, 对各国各项业务的顺利进行起到不可或缺的作用, 所以当其发生危机时, 会直接对全球范围内的金融机构造成负面影响. 学术界对如何识别中国系统重要性银行进行了很多有益尝试, 由于研究方法或样本不同, 得出的结论存在一定差异. 有效识别此类银行是当前的热点议题. 文章从系统重要性银行的度量数据出发, 首先以各银行的财务报表数据和股票价格数据为研究样本. 其次, 在~SVM-Copula 集成系统基础上, 利用粒子群优化算法对~SVM 寻找最优参数组合, 进而提出了优于~GARCH 模型和核密度估计法的~PSO-SVM 边缘分布估计法. 最后将~PSO-SVM-Copula 集成系统运用到~CoVaR 领域中. 研究结果表明: PSO-SVM-Copula-CoVaR (PSCC) 模型在系统重要性银行的评估上比仅使用单方面的数据更加合理.
Systemically important banks is the junction of the global business chain, which plays an indispensable role in the smooth operation of various countries' businesses. So when the crisis occurs, will directly have a negative impact on global financial institutions. Academic circles made lot of beneficial attempts on identify systemically important Banks in China. Due to different research methods or samples, the conclusions are different. Effectively identify such bank is the current hot issues. This paper starts with the measured data of the systemically important banks. Firstly, take the financial statements and stock price data of each bank as the research samples. Secondly, based on the SVM-Copula integrated system, use the particle swarm optimization algorithm to find the optimal parameter combination for SVM. Furthermore, this paper proposes the PSO-SVM edge distribution estimation method which is superior to GARCH model and kernel density estimation method. Finally, apply the PSO-SVM-Copula integrated system to the CoVaR field. The results show that the PSO-SVM-Copula-CoVaR (PSCC) model is more reasonable in evaluating systemically important banks than using only unilateral data.
系统重要性银行 / 系统性风险 / 支持向量机 / 粒子群优化算法 / Copula-CoVaR. {{custom_keyword}} /
/
〈 | 〉 |