Does Dataset Complexity Matters for Model Explainers?

TL;DR

This paper evaluates how dataset complexity influences the effectiveness of various XAI methods in ranking feature importance across multiple models and tools.

Contribution

It provides an experimental benchmark analyzing the impact of data complexity on the consistency of explainability measures in tabular data.

Findings

Explainability measures often produce different rankings.

Dataset complexity affects the stability of model explanations.

Benchmark results highlight the importance of data characteristics in XAI.

Abstract

Strategies based on Explainable Artificial Intelligence - XAI have emerged in computing to promote a better understanding of predictions made by black box models. Most XAI measures used today explain these types of models, generating attribute rankings aimed at explaining the model, that is, the analysis of Attribute Importance of Model. There is no consensus on which XAI measure generates an overall explainability rank. For this reason, several proposals for tools have emerged (Ciu, Dalex, Eli5, Lofo, Shap and Skater). An experimental benchmark of explainable AI techniques capable of producing global explainability ranks based on tabular data related to different problems and ensemble models are presented herein. Seeking to answer questions such as "Are the explanations generated by the different measures the same, similar or different?" and "How does data complexity play along model explainability?" The results from the construction of 82 computational models and 592 ranks shed some light on the other side of the problem of explainability: dataset complexity!