A New Common Weights DEA Model Based on Cluster Analysis

Nam Hyok Kim,Feng He,Kwon Ryong Hong,Hyok-Chol Kim,Sok-Min Han
DOI: https://doi.org/10.1007/s12351-024-00838-5
IF: 2.7
2024-01-01
Operational Research
Abstract:The data envelopment analysis (DEA) is a data-driven tool for performance evaluation. Standard DEA assigns the most favorable weights to decision-making units (DMUs), so it is impossible to compare and rank those on the same basis. Common weights DEA models assign a common weight vector to all DMUs to provide the same basis for evaluation, and most studies have been discussed only in terms of the distance between an ideal value and a real value. The paper proposes a new common weights DEA model based on cluster analysis. A clustering method of DMUs by the production function is suggested by using the global reference set and the determination of the common weights is discussed. The numerical experiments are illustrated to examine the validity of the proposed model, and the experiments show that the model gives reasonable results compared to previous studies. The common weights DEA model is applied to evaluate the environmental efficiency of China’s 48 iron and steel enterprises. The proposed model is the first study for obtaining common weights by considering cluster analysis.
What problem does this paper attempt to address?