Skip to main content
QUICK REVIEW

[论文解读] Portfolio Optimization based on Neural Networks Sensitivities from Assets Dynamics respect Common Drivers

Alejandro Rodríguez Domínguez|arXiv (Cornell University)|Feb 17, 2022
Financial Markets and Investment Strategies被引用 5
一句话总结

本文提出了一种新颖的组合优化框架,通过偏微分方程(PDEs)建模资产动态,并利用神经网络近似其解,通过自动伴随微分(AAD)计算敏感度。该框架提出共性原则以识别最优的共同驱动因素,将资产嵌入敏感度空间,并对敏感度矩阵进行层次聚类,从而在特有风险与系统性风险方面实现更优的分散化,相较于现有样本外方法在多个市场中表现更优。

ABSTRACT

We present a framework for modeling asset and portfolio dynamics, incorporating this information into portfolio optimization. We define drivers for asset and portfolio dynamics and their optimal selection. For this framework, we introduce the Commonality Principle, providing a solution for the optimal selection of portfolio drivers as the common drivers. Portfolio constituent dynamics are modeled by Partial Differential Equations, and solutions approximated with neural networks. Sensitivities with respect to the common drivers are obtained via Automatic Adjoint Differentiation. Information on asset dynamics is incorporated via sensitivities into portfolio optimization. Portfolio constituents are embedded into the space of sensitivities with respect to their common drivers, and a distance matrix in this space called the Sensitivity matrix is used to solve the convex optimization for diversification. The sensitivity matrix measures the similarity of the projections of portfolio constituents on a vector space formed by common drivers’ returns and is used to optimize for diversification on both idiosyncratic and systematic risks while adding directionality and future behavior information via returns dynamics. For portfolio optimization, we perform hierarchical clustering on the sensitivity matrix. The clustering tree is used for recursive bisection to obtain the weights. To the best of the author’s knowledge, this is the first time that sensitivities’ dynamics approximated with neural networks have been used for portfolio optimization. Secondly, that hierarchical clustering on a matrix of sensitivities is used to solve the convex optimization problem and incorporate the hierarchical information of these sensitivities. Thirdly, public and listed variables can be used to obtain maximum idiosyncratic and systematic diversification by means of the sensitivity space with respect to optimal portfolio drivers. We reach over-performance in many experiments with respect to all other out-of-sample methods for different markets and real datasets. We also include a recipe for the methodology to increase performance even further, and tackle the main issues in portfolio management such as regimes, non-stationarity, overfitting, and selection bias.

研究动机与目标

  • 开发一种将资产与组合的动态行为纳入组合优化框架的方法,结合神经网络与敏感度分析。
  • 基于因果关系、持续性与不同时滞下的相关性,识别组合构成资产的最优共同驱动因素。
  • 通过将资产相对于共同驱动因素嵌入敏感度空间,实现最大化的分散化,同时捕捉特有风险与系统性风险。
  • 通过利用动态敏感度而非静态相关性,改进传统的均值-方差与层次风险平价(HRP)方法。
  • 通过数据驱动的、基于动态敏感度的方法,解决组合管理中的关键挑战,如非平稳性、过拟合、状态转换与选择偏差。

提出的方法

  • 利用通过共性原则识别的共同外生因素,建立基于偏微分方程(PDEs)的资产动态模型。
  • 通过在时间序列数据上训练的神经网络,近似这些PDE的解,并计算相对于共同驱动因素的敏感度。
  • 通过自动伴随微分(AAD)计算敏感度,以获得资产收益对驱动因素的精确、可微分梯度。
  • 将各资产的敏感度值聚合为向量空间,形成捕捉驱动因素空间中资产间几何关系的敏感度矩阵。
  • 对敏感度矩阵应用层次聚类,提取层次结构,并通过递归二分法实现组合权重的凸优化。
  • 利用所得的层次敏感度平价(HSP)方法优化分散化,整合来自公开与上市数据的未来行为与收益动态。

实验结果

研究问题

  • RQ1基于神经网络近似的资产收益对共同驱动因素的敏感度,是否能超越传统基于相关性的方法,实现更优的组合分散化?
  • RQ2如何选择最优的共同驱动因素集合,以在不同时间滞后下实现最大持久性、因果关系与相关性?
  • RQ3在组合优化中,对敏感度矩阵进行层次聚类在多大程度上优于对相关性矩阵进行层次聚类?
  • RQ4来自PDE解的动态敏感度信息是否能提升在多样化市场与市场状态下的样本外组合表现?
  • RQ5该框架如何通过使用现实世界中的上市变量,利用算法筛选过程有效缓解过拟合、选择偏差与非平稳性问题?

主要发现

  • 所提出的层次敏感度平价(HSP)方法在多个市场与真实数据集(包括股票指数、货币与因子ETF)中,持续优于所有其他样本外方法。
  • 由神经网络近似PDE解与AAD生成的敏感度矩阵,相较于传统相关性矩阵,能捕捉更丰富的动态风险与收益信息。
  • 将敏感度投影中的层次结构纳入考虑,可实现优于仅依赖资产收益相关性或余弦相似度的分散化效果。
  • 共性原则成功识别出基于时滞间持久且高相关性关系的最优共同驱动因素,提升了模型的稳健性与预测能力。
  • 通过经过调优的算法筛选过程,该框架有效缓解了驱动因素选择中的过拟合与选择偏差,过滤了虚假相关性与多重共线性。
  • 该方法仅使用公开与上市变量作为驱动因素,即可实现特有风险与系统性风险的最大化分散化,无需依赖统计因子或风险溢价模型。

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。