Data Envelopment Analysis in Business and Economics: Methods, Trends, and Research Frontiers

Abstract

Data Envelopment Analysis (DEA) has become an important multidimensional research method and is now widely used in business and economics. This study adopts a scoping‐review approach and uses Natural Language Processing (NLP), a data‐driven machine‐learning technique, to analyze 12,686 relevant studies and classify the literature into major research themes and categories. Specifically, we identify three DEA‐related themes, CoreDEA (11,167 studies), StatDEA (1266 studies), and DecisionDEA (253 studies), which can be further divided into 15 categories, including Dynamic DEA, Undesirable DEA, Bootstrap DEA, and ANP/AHP DEA. The findings show that traditional DEA methods, especially those within CoreDEA, continue to develop actively. At the same time, the growing prominence of sustainability‐related research suggests that SBM DEA and hybrid models incorporating bootstrap, meta‐frontier, and network DEA may represent an important direction for future research.