Understanding Interventional TreeSHAP : How and Why it Works - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2022

Understanding Interventional TreeSHAP : How and Why it Works

Gabriel Laberge
Yann Batiste Pequignot

Résumé

Shapley values are ubiquitous in interpretable Machine Learning due to their strong theoretical background and efficient implementation in the SHAP library. Computing these values previously induced an exponential cost with respect to the number of input features of an opaque model. Now, with efficient implementations such as Interventional TreeSHAP, this exponential burden is alleviated assuming one is explaining ensembles of decision trees. Although Interventional TreeSHAP has risen in popularity, it still lacks a formal proof of how/why it works. We provide such proof with the aim of not only increasing the transparency of the algorithm but also to encourage further development of these ideas. Notably, our proof for Interventional TreeSHAP is easily adapted to Shapley-Taylor indices and one-hot-encoded features.
Fichier principal
Vignette du fichier
2209.15123.pdf (296.89 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04232220 , version 1 (07-10-2023)

Identifiants

Citer

Gabriel Laberge, Yann Batiste Pequignot. Understanding Interventional TreeSHAP : How and Why it Works. 2023. ⟨hal-04232220⟩
22 Consultations
140 Téléchargements

Altmetric

Partager

Gmail Mastodon Facebook X LinkedIn More