Exploring XAI techniques for enhancing model transparency and interpretability in real estate rent prediction: A comparative study

Abstract

Black-box artificial intelligence (AI) models are popular in real estate research, but their lack of interpretability raises concerns. To address this, explainable AI (XAI) techniques have been applied to shed light on these models. This paper presents a comparative study of six global XAI techniques on a CatBoost model for Belgian residential rent prediction. Results show that while some techniques offer substitute insights, others provide complementary perspectives on the model's behavior. Employing multiple XAI techniques is crucial to comprehensively understand rents drivers which contributes to transparency, interpretability, and model governance in the real estate industry, advancing the adoption of (X)AI.