基于蒙特卡洛小波去噪的股票投资组合风险优化研究  

Stock portfolio risk optimization based on filtering method of wavelet using Monte Carlo simulation

作　　者：李君昌 樊重俊[1] 杨云鹏[1] 袁光辉 Li Junchang;Fan Chongjun;Yang Yunpeng;Yuan Guanghui(Business School,University of Shanghai for Science&Technology,Shanghai 200093,China;School of Information Management&Engineering,Shanghai University of Finance&Economics,Shanghai 200433,China;Experimental Center,Shanghai University of Finance&Economics,Shanghai 200433,China)

机构地区：[1]上海理工大学管理学院,上海200093 [2]上海财经大学,信息管理与工程学院,上海200433 [3]上海财经大学,实验中心,上海200433

出　　处：《计算机应用研究》2018年第10期2947-2951,3028,共6页Application Research of Computers

基　　金：国家自然科学基金资助项目(71303157);上海市教育委员会科研创新重点基金资助项目(14ZZ131);上海市一流学科基金资助项目(S1205YLXK);上海市社科规划青年课题基金资助项目(2014EGL007);沪江基金资助项目(D14008)

摘　　要：股票投资是我国重要的投资形式之一,但是从股票市场到投资者的个人投资决策都充斥着干扰和噪声。传统的随机矩阵理论(RMT)对证券投资组合风险优化产生了显著效果,然而RMT的去噪效果会随着组合证券数量的减少而下降,同时RMT仅仅是对证券收益协方差矩阵进行去噪,所以存在着一定局限性。从马科维茨(Markowitz)投资组合风险优化模型出发,基于小波分析视角讨论了蒙特卡洛模拟与小波去噪应用在投资组合风险优化上的可行性,并构造了蒙特卡洛小波去噪算法(MWF)。运用上海证券交易所A股股票交易数据进行实证分析,对比了传统的蒙特卡洛RMT(MKR)去噪方法、传统的RMT去噪法(LCPB、PG+和KR)以及单纯的Markowitz模型在股票投资组合上的风险优化效果。实证结果表明,蒙特卡洛小波去噪法在股票投资组合风险优化方面表现更好,并提供了一种新的股票投资组合配置方式。Stock investment is one of the most important investment forms in China.Nevertheless,it is filled with interference and noise in the procedure from stock’s market to individual investment decision.The traditional method,random matrix theory(RMT),plays a critical role in risk optimization of stock portfolio,but the de-noising effect of RMT is deteriorating as the number of portfolio reducing.At the same time,RMT only can filter on securities revenue covariance matrix.Consequently,to some extent,there are many limitations.This paper discussed the feasibility of Monte Carlo simulation and wavelet filtering that applied in investment risk optimization from the Markowitz portfolio risk optimization model,based on wavelet analysis perspective.Then it designed the filtering method of wavelet using Monte Carlo simulation(MWF).By using the A shares stock exchange data of Shanghai stock exchange,it conducted the empirical analysis to compared with the risk optimization effect of traditional Monte Carlo RMT(MKR)filtering method,traditional RMT filtering method(LCPB,PG+and KR)and simple Markowitz model in the stock investment.The empirical result shows that the filtering method of wavelet using Monte Carlo simulation performs better and provides a new configuration method in terms of stock portfolio risk optimization.

关 键 词：马科维茨 投资组合 蒙特卡洛模拟 小波去噪 风险优化 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] F830.59[自动化与计算机技术—计算机科学与技术]