摘要
在大数据时代,高维资产对于很多金融机构非常常见,维数诅咒的影响使得在投资组合中扮演着重要角色的协方差阵的估计效率较低。将子集重采样方法应用到投资组合中,首先,从所有资产构造的集合中抽取若干个子集;然后,采用BEKK模型来估计和预测子集的协方差阵,以解决维数诅咒问题;最后,对若干个子集的同一个资产的权重向量求平均,来求得每个资产的权重。通过实证分析发现,基于子集重抽样的投资组合明显要优于传统的均值——方差投资组合,其收益更高、波动更小,并且夏普比率值也较高。
In the era of big data,high dimensional assets are very common for many financial institutions. Because of the curse of the dimensionality,the estimation efficiency of covariance matrix which plays an import role in portfolio is low. This paper applies the subset resampling to portfolio. First,several subsets are extracted from the set of all assets;Then the BEKK model is used to estimate and predict the covariance matrix of subsets,to solve the curse of dimensionality. Finally,the weights of the same assets of several subsets are averaged to obtain the weight of each asset. Through empirical studies,we find that：the portfolio which based on subset resampling is significantly better than the mean-variance portfolio,with higher returns,smaller volatility,and higher Sharpe ratio.
作者
刘丽萍
LIU Li-ping(School of Mathematics and Statistics,Guizhou University of Finance and Economics,Guiyang 550025,Chin)
出处
《经济研究导刊》
2018年第19期78-81,共4页
Economic Research Guide
基金
国家社会科学基金项目(16CTJ013)
