Evaluation Metrics for Data Valuation Methods

Data Valuation has emerged as a critical research topic at the intersection of artificial intelligence and economics, enabling principled quantification of individual data importance. Despite many proposed algorithms, a central challenge remains: how to evaluate their quality, reliability, and fairness. This paper presents the first systematic review of evaluation metrics for data valuation, synthesising perspectives from data science and economics. We categorise existing metrics into four data science-oriented classes and complement them with economic principles and practical guidance for metric selection. This cross-disciplinary framework promotes more rigorous and standardised evaluation protocols in future data valuation research.