From Economic Explanation to Performance Improvement in Machine-Learning Portfolios

By Yunqi Liang, Stefan Zohren, Mihai Cucuringu

Rating

1817
Battle Count: 53

Relevance

10/10
Highly relevant. The paper directly addresses the gap between model interpretability and economic performance in ML-driven portfolios, offering a practical method to improve net Sharpe ratios via post-training feature scaling without retraining.

Implementation Complexity

7/10
Requires implementing SHAP values for specific models (TreeSHAP, DeepSHAP), deriving analytical gradients for portfolio metrics (turnover, Sharpe ratio) with respect to feature exposures, and managing the temporal stabilization logic. The mathematical derivation is complex, but the application is modular.

Reproducibility

3/5
The paper provides detailed mathematical formulations, algorithm descriptions, and hyperparameter settings. It uses public data (Kenneth French Data Library, CRSP/Compustat implied by US equity data). However, no explicit GitHub repository URL is provided in the text extract.

Abstract

To extend conventional model-level feature importance to economically meaningful portfolio outcomes and decisions, we propose a local first-order attribution-propagation framework that maps allocation-function feature exposures to local sensitivities of sequential portfolio performance metrics. We relate the propagated attributions to Shapley values defined directly for portfolio objectives and derive statistical inference for Sharpe-ratio attributions. The resulting attribution-guided feature-exposure intervention enables post-training adaptation, translating portfolio-metric explanations into out-of-sample decisions without retraining the underlying allocation model. We further introduce temporal stabilisation and cross-model averaging as complementary regularisation mechanisms. Empirically, the interventions yield statistically significant improvements in net Sharpe ratios across different model classes.

Links

arXiv:2610.15144 ยท Paper rankings

The interactive Everscope explorer (charts, battles, favorites) loads below.