Aggregate Models, Not Explanations: Improving Feature Importance Estimation

TL;DR

This paper demonstrates that aggregating ensemble models yields more accurate feature importance estimates than explaining individual models, especially for complex models, supported by theoretical analysis and empirical validation.

Contribution

It provides a theoretical framework showing model-level ensembling improves feature importance estimation over individual explanations for complex models.

Findings

Model ensembling reduces importance estimation error.

Ensembling at the model level outperforms individual explanations.

Validated on benchmarks and large-scale biomedical data.

Abstract

Feature-importance methods show promise in transforming machine learning models from predictive engines into tools for scientific discovery. However, due to data sampling and algorithmic stochasticity, expressive models can be unstable, leading to inaccurate variable importance estimates and undermining their utility in critical biomedical applications. Although ensembling offers a solution, deciding whether to explain a single ensemble model or aggregate individual model explanations is difficult due to the nonlinearity of importance measures and remains largely understudied. Our theoretical analysis, developed under assumptions accommodating complex state-of-the-art ML models, reveals that this choice is primarily driven by the model's excess risk. In contrast to prior literature, we show that ensembling at the model level provides more accurate variable-importance estimates, particularly for expressive models, by reducing this leading error term. We validate these findings on classical benchmarks and a large-scale proteomic study from the UK Biobank.