Consistent feature attribution for tree ensembles

Scott Lundberg, Su-In Lee

Jun 16, 2017 (modified: Jun 19, 2017) ICML 2017 WHI Submission readers: everyone
  • Abstract: It is critical in many applications to understand what features are important for a model, and why individual predictions were made. For tree ensemble methods these questions are usually answered by attributing importance values to input features, either globally or for a single prediction. Here we show that current feature attribution methods are inconsistent, which means changing the model to rely more on a given feature can actually decrease the importance assigned to that feature. To address this problem we develop fast exact solutions for SHAP (SHapley Additive exPlanation) values, which were recently shown to be the unique additive feature attribution method based on conditional expectations that is both consistent and locally accurate. We integrate these improvements into the latest version of XGBoost, demonstrate the inconsistencies of current methods, and show how using SHAP values results in significantly improved supervised clustering performance. Feature importance values are a key part of understanding widely used models such as gradient boosting trees and random forests. We believe our work improves on the state-of-the-art in important ways, and so impacts any current user of tree ensemble methods.
  • TL;DR: Current methods for interpreting tree ensembles are inconsistent, we present fast algorithms for SHAP values to address this problem.
  • Keywords: SHAP values, tree ensembles, feature attribution, supervised clustering, Shapley

Loading