Saved in:
Bibliographic Details
Main Authors: Ferrere, Baptiste, Bousquet, Nicolas, Gamboa, Fabrice, Loubes, Jean-Michel, Muré, Joseph
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.02673
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912939309531136
author Ferrere, Baptiste
Bousquet, Nicolas
Gamboa, Fabrice
Loubes, Jean-Michel
Muré, Joseph
author_facet Ferrere, Baptiste
Bousquet, Nicolas
Gamboa, Fabrice
Loubes, Jean-Michel
Muré, Joseph
contents Functional ANOVA offers a principled framework for interpretability by decomposing a model's prediction into main effects and higher-order interactions. For independent features, this decomposition is well-defined, strongly linked with SHAP values, and serves as a cornerstone of additive explainability. However, the lack of an explicit closed-form expression for general dependent distributions has forced practitioners to rely on costly sampling-based approximations. We completely resolve this limitation for categorical inputs. By bridging functional analysis with the extension of discrete Fourier analysis, we derive a closed-form decomposition without any assumption. Our formulation is computationally very efficient. It seamlessly recovers the classical independent case and extends to arbitrary dependence structures, including distributions with non-rectangular support. Furthermore, leveraging the intrinsic link between SHAP and ANOVA under independence, our framework yields a natural generalization of SHAP values for the general categorical setting.
format Preprint
id arxiv_https___arxiv_org_abs_2603_02673
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Exact Functional ANOVA Decomposition for Categorical Inputs Models
Ferrere, Baptiste
Bousquet, Nicolas
Gamboa, Fabrice
Loubes, Jean-Michel
Muré, Joseph
Machine Learning
Functional ANOVA offers a principled framework for interpretability by decomposing a model's prediction into main effects and higher-order interactions. For independent features, this decomposition is well-defined, strongly linked with SHAP values, and serves as a cornerstone of additive explainability. However, the lack of an explicit closed-form expression for general dependent distributions has forced practitioners to rely on costly sampling-based approximations. We completely resolve this limitation for categorical inputs. By bridging functional analysis with the extension of discrete Fourier analysis, we derive a closed-form decomposition without any assumption. Our formulation is computationally very efficient. It seamlessly recovers the classical independent case and extends to arbitrary dependence structures, including distributions with non-rectangular support. Furthermore, leveraging the intrinsic link between SHAP and ANOVA under independence, our framework yields a natural generalization of SHAP values for the general categorical setting.
title Exact Functional ANOVA Decomposition for Categorical Inputs Models
topic Machine Learning
url https://arxiv.org/abs/2603.02673