Luisa Bouneder
X-SHAP: towards multiplicative explainability of Machine Learning

This paper introduces X-SHAP, a model-agnostic method that assesses multiplicative contributions of variables for both local and global predictions.

This methodtheoretically and operationally extends the so-called additive SHAP approach. Itproves useful underlying multiplicative interactions of factors, typically arisingin sectors where Generalized Linear Models are traditionally used, such as ininsurance or biology.

We test the method on various datasets and propose a set oftechniques based on individual X-SHAP contributions to build aggregated multiplicative contributions and to capture multiplicative feature importance, that wecompare to traditional techniques.