Explaining predictive models using Shapley values and non-parametric vine copulas

Dependence Modeling, vol. 9, pp. 62–81

Abstract

Abstract In this paper the goal is to explain predictions from complex machine learning models. One method that has become very popular during the last few years is Shapley values. The original development of Shapley values for prediction explanation relied on the assumption that the features being described were independent. If the features in reality are dependent this may lead to incorrect explanations. Hence, there have recently been attempts of appropriately modelling/estimating the dependence between the features. Although the previously proposed methods clearly outperform the traditional approach assuming independence, they have their weaknesses. In this paper we propose two new approaches for modelling the dependence between the features. Both approaches are based on vine copulas, which are flexible tools for modelling multivariate non-Gaussian distributions able to characterise a wide range of complex dependencies. The performance of the proposed methods is evaluated on simulated data sets and a real data set. The experiments demonstrate that the vine copula approaches give more accurate approximations to the true Shapley values than their competitors.

Authors 4

1. Kjersti Aas corresponding

   Norwegian Computing Center

   Affiliation as printed

   Norwegian Computing Center

2. Thomas Nagler Aachen

   Leiden University

   Affiliation as printed

   Leiden University

3. Martin Jullum

   Norwegian Computing Center

   Affiliation as printed

   Norwegian Computing Center

4. Anders Løland

   Norwegian Computing Center

   Affiliation as printed

   Norwegian Computing Center