基于可信性均值-方差-偏度-正弦熵的投资组合模型

A mean-variance-skewness-sine entropy model for portfolio selection based on credibility theory

通过引入可信性理论和偏度约束,分别建立了同时满足随机不确定和模糊不确定情形的均值-方差-偏度-正弦熵多目标投资组合优化模型(Mult-M-V-S-SE)和含有收益及风险系数的单目标投资组合优化模型(M-V-SSE);然后运用马尔科夫方法预测模糊收益率,利用遗传算法优化模型投资策略,通过上海证券交易所数据进行实证比较。结果表明:Mult-M-V-S-SE和M-V-S-SE两模型效果均超过均匀投资策略(MEAN);M-V-S-SE模型灵活且稳定,更具有优势,在满足投资者需求的同时可实现更高的累计收益。

By introducing the credibility theory and skewness,a multi-objective mean-variance-skewness-sine entropy（mult-M-V-S-SE） portfolio selection model considering random uncertainty and fuzzy uncertainty and a single objective portfolio selection model containing return and risk factors are proposed.The Markov method is employed to forecast the fuzzy yield,and genetic algorithms are introduced in order to obtain the optimal portfolio.An empirical comparison was conducted by employing the 2015 stock price data from the Shanghai stock exchange.The results illustrate that the two models perform better than the average investment strategy,and the single objective model is more flexible and stable,which will afford higher cumulative yield rates which meet the needs of investors.