Inferring feature importance with uncertainties in high-dimensional data
This work addresses the need for reliable feature importance explanations in data-based models, particularly for high-dimensional domains like genomics, but it is incremental as it builds upon existing SAGE methods.
The paper tackles the problem of estimating feature importance with uncertainties in high-dimensional data, introducing a Shapley value-based framework called sub-SAGE that can be estimated without resampling for tree-based models, as demonstrated on synthetic and genomics data.
Estimating feature importance is a significant aspect of explaining data-based models. Besides explaining the model itself, an equally relevant question is which features are important in the underlying data generating process. We present a Shapley value based framework for inferring the importance of individual features, including uncertainty in the estimator. We build upon the recently published feature importance measure of SAGE (Shapley additive global importance) and introduce sub-SAGE which can be estimated without resampling for tree-based models. We argue that the uncertainties can be estimated from bootstrapping and demonstrate the approach for tree ensemble methods. The framework is exemplified on synthetic data as well as high-dimensional genomics data.