Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles
arXiv:2401.12069 · doi:10.1609/aaai.v38i13.29352
Abstract
While shallow decision trees may be interpretable, larger ensemble models like gradient-boosted trees, which often set the state of the art in machine learning problems involving tabular data, still remain black box models. As a remedy, the Shapley value (SV) is a well-known concept in explainable artificial intelligence (XAI) research for quantifying additive feature attributions of predictions. The model-specific TreeSHAP methodology solves the exponential complexity for retrieving exact SVs from tree-based models. Expanding beyond individual feature attribution, Shapley interactions reveal the impact of intricate feature interactions of any order. In this work, we present TreeSHAP-IQ, an efficient method to compute any-order additive Shapley interactions for predictions of tree-based models. TreeSHAP-IQ is supported by a mathematical framework that exploits polynomial arithmetic to compute the interaction scores in a single recursive traversal of the tree, akin to Linear TreeSHAP. We apply TreeSHAP-IQ on state-of-the-art tree ensembles and explore interactions on well-established benchmark datasets.
References in corpus (13)
- Scikit-learn: Machine Learning in Python
- XGBoost: A Scalable Tree Boosting System
- Understanding Global Feature Contributions With Additive Importance Measures
- Fooling LIME and SHAP: Adversarial Attacks on Post hoc Explanation Methods
- Fast TreeSHAP: Accelerating SHAP Value Computation for Trees
- How does this interaction affect me? Interpretable attribution for feature interactions
- Faith-Shap: The Faithful Shapley Interaction Index
- From Shapley Values to Generalized Additive Models and back
- SHAP-IQ: Unified Approximation of any-order Shapley Interactions
- Purifying Interaction Effects with the Functional ANOVA: An Efficient Algorithm for Recovering Identifiable Additive Models
- Unifying local and global model explanations by functional decomposition of low dimensional structures
- High Dimensional Model Explanations: an Axiomatic Approach
- Decomposing Global Feature Effects Based on Feature Interactions