Large-scale Time-Varying Portfolio Optimisation using Graph Attention Networks

2024-07-22

arXiv.org

AI Summary

This paper addresses the dynamic portfolio optimization problem for firms exposed to delisting and default risk. Methodologically, it introduces the first end-to-end graph learning framework that jointly models nonlinear asset dependencies and maximizes risk-adjusted returns. It constructs time-varying asset relational graphs using distance-based correlation and the Thresholded Markov Random Field Graph (TMFG), employs a Graph Attention Network (GAT) to jointly capture topological dynamics and Sharpe ratio maximization, and incorporates an interpretable weight-constrained layer. Contributions include: (i) the first integration of high-risk firms into large-scale portfolio optimization; (ii) a novel Sharpe-ratio-driven end-to-end GAT training paradigm; and (iii) unified topology-awareness, constraint interpretability, and performance optimizability. Empirical evaluation on a 30-year mid-cap stock dataset demonstrates significant outperformance over mean-variance, network-feature-based, and equal-weight benchmarks—yielding robust alpha and heightened sensitivity to structural market shifts.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Search and Optimization: Non-convex Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies

Abstract

Apart from assessing individual asset performance, investors in financial markets also need to consider how a set of firms performs collectively as a portfolio. Whereas traditional Markowitz-based mean-variance portfolios are widespread, network-based optimisation techniques offer a more flexible tool to capture complex interdependencies between asset values. However, most of the existing studies do not contain firms at risk of default and remove any firms that drop off indices over a certain time. This is the first study to also incorporate such firms in portfolio optimisation on a large scale. We propose and empirically test a novel method that leverages Graph Attention networks (GATs), a subclass of Graph Neural Networks (GNNs). GNNs, as deep learning-based models, can exploit network data to uncover nonlinear relationships. Their ability to handle high-dimensional data and accommodate customised layers for specific purposes makes them appealing for large-scale problems such as mid- and small-cap portfolio optimisation. This study utilises 30 years of data on mid-cap firms, creating graphs of firms using distance correlation and the Triangulated Maximally Filtered Graph approach. These graphs are the inputs to a GAT model incorporating weight and allocation constraints and a loss function derived from the Sharpe ratio, thus focusing on maximising portfolio risk-adjusted returns. This new model is benchmarked against a network characteristic-based portfolio, a mean variance-based portfolio, and an equal-weighted portfolio. The results show that the portfolio produced by the GAT-based model outperforms all benchmarks and is consistently superior to other strategies over a long period, while also being informative of market dynamics.

Problem

• Optimize large-scale time-varying portfolios
• Incorporate firms at risk of default
• Utilize Graph Attention Networks for optimization

Innovation

• Utilizes Graph Attention Networks
• Incorporates firms at risk
• Maximizes risk-adjusted returns