Tractable Explaining of Multivariate Decision Trees - Argumentation, Décision, Raisonnement, Incertitude et Apprentissage Access content directly
Conference Papers Year : 2023

Tractable Explaining of Multivariate Decision Trees

Abstract

We study multivariate decision trees (MDTs), in particular, classes of MDTs determined by the language of relations that can be used to split feature space. An abductive explanation (AXp) of the classification of a particular instance, viewed as a set of feature-value assignments, is a minimal subset of the instance which is sufficient to lead to the same decision. We investigate when finding a single AXp is tractable. We identify tractable languages for real, integer and boolean features. Indeed, in the case of boolean languages, we provide a P/NP-hard dichotomy.
Fichier principal
Vignette du fichier
tractableexplaining.pdf (289.82 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04268587 , version 1 (02-11-2023)

Identifiers

Cite

Clément Carbonnel, Martin Cooper, Joao Marques-Silva. Tractable Explaining of Multivariate Decision Trees. KR 2023 - 20th International Conference on Principles of Knowledge Representation and Reasoning, Sep 2023, Rhodes, Greece. pp.127-135, ⟨10.24963/kr.2023/13⟩. ⟨hal-04268587⟩
127 View
68 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More