Explaining a probabilistic prediction on the simplex with Shapley compositions - Archive ouverte HAL
Communication Dans Un Congrès Année : 2024

Explaining a probabilistic prediction on the simplex with Shapley compositions

Résumé

Originating in game theory, Shapley values are widely used for explaining a machine learning model's prediction by quantifying the contribution of each feature's value to the prediction. This requires a scalar prediction as in binary classification, whereas a multiclass probabilistic prediction is a discrete probability distribution, living on a multidimensional simplex. In such a multiclass setting the Shapley values are typically computed separately on each class in a one-vs-rest manner, ignoring the compositional nature of the output distribution. In this paper, we introduce Shapley compositions as a well-founded way to properly explain a multiclass probabilistic prediction, using the Aitchison geometry from compositional data analysis. We prove that the Shapley composition is the unique quantity satisfying linearity, symmetry and efficiency on the Aitchison simplex, extending the corresponding axiomatic properties of the standard Shapley value. We demonstrate this proper multiclass treatment in a range of scenarios.
Fichier principal
Vignette du fichier
Explaining_a_probabilistic_prediction_on_the_simplexwith_Shapley_compositions.pdf (722.94 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04687356 , version 1 (04-09-2024)

Identifiants

Citer

Paul-Gauthier Noé, Miquel Perelló-Nieto, Jean-François Bonastre, Peter Flach. Explaining a probabilistic prediction on the simplex with Shapley compositions. ECAI 2024, Oct 2024, Santiago de Compostela, Spain. ⟨hal-04687356⟩

Collections

UNIV-AVIGNON LIA
3 Consultations
2 Téléchargements

Altmetric

Partager

More