基于D-vinecopula-分位数回归的组合投资决策.PDF

第34 卷第1 期 系统工 程 学 报 Vol.34 No.1 2019 年2 月 JOURNAL OF SYSTEMS ENGINEERING Feb. 2019 基于D-vine copula-分位数回归的组合投资决策 , , 许启发 , 王侠英 , 蒋翠侠 , 李辉艳 (1. 合肥工业大学管理学院, 安徽合肥230009; 2. 合肥工业大学过程优化与智能决策教育部重点实验室, 安徽合肥230009) 摘要: 为克服传统组合投资决策模型使用方差风险的不足, 建立D-vine copula-分位数回归方法估计多元条件联合 分布, 给出广义Omega 比率组合投资决策模型求解方案. 分别选取能源市场3种期货商品和不同行业5只股票进行实 证研究, 结果表明: 基于D-vine copula-分位数回归的广义Omega 比率组合投资决策模型, 能够充分揭示与模拟金融 资产收益变动规律, 得到更高的Sharpe 比率和广义Omega 比率. 关键词: 组合投资; D-vine copula; 分位数回归; 广义Omega 比率 中图分类号: F224.0 文献标识码: A 文章编号:2019 doi: 10.13383/ki.jse.2019.01.006 Portfolio selection via D-vine copula-quantile regression method Xu Qifa , Wang Xiaying , Jiang Cuixia , Li Huiyan (1. School of Management, Hefei University of Technology, Hefei 230009, China; 2. Key Laboratory of Process Optimization and Intelligent Decision-making, Ministry of Education, Hefei 230009, China) Abstract: In classical portfolio selection models, variance is an important parameter. However, it is not suitable for risk measure. To address this issue, this paper develops a generalized Omega ratio-based portfolio selec- tion model by using a D-vine copula-quantile regression to estimate complete joint probability distribution of portfolio returns. The efficacy of the new method is illustrated through empirical studies on global energy markets and Chinese stock market. Two practical portfolios are constructed on three kinds of commodities and five stocks respectively. The empirical re