论文部分内容阅读
给出了风险条件下基于收益视角的损失最小化投资组合模型.数值试验表明,当投资者预期收益较高时,该模型等价于条件在险价值模型,当投资者满足于较低收益时,该模型优于条件在险价值模型.提出的基于主成分分析的情景生成方法避免了过度分散投资,投资组合模型的最优目标值随情景数目的增加快速趋于收敛;该情景生成方法无需假定随机收益的分布,融合了收益分布的非对称及尖峰等统计特征,主成分分析法的降维优势使该方法也适合于资产数目较大时的情景生成.
The loss minimization portfolio model based on profit perspective is given under the risk condition.Numerical tests show that when the expected return of investors is high, the model is equivalent to the conditional VaR model, and when investors are satisfied with the lower returns , Which is superior to the condition-based VaR model.The proposed scenario generation method based on principal component analysis avoids the excessive diversification investment and the optimal target value of the portfolio model converges quickly with the increase of the number of scenarios. This scenario generation method does not need Assuming the distribution of random benefits, the statistical characteristics such as asymmetry and spike of income distribution are integrated. The advantage of dimensionality reduction of principal component analysis makes this method suitable for scenario generation when the number of assets is large.