Find Role Models Through a Social Network Data Envelopment Analysis Method and its Application on Chunyu Doctor Platform

Identifying role models from a large-scale sample is critical for improving the performance of the society or large organizations because it is an effective way to spread positive ideologies, attitudes, and the best behaviors. To find role models from a large-scale sample, this study proposes a social network data envelopment analysis method by integrating data envelopment analysis (DEA) and social network analysis (SNA), where DEA is non-parametric data-driven technique for evaluating the relative performance of peer decision making units (DMUs) and SNA seeks to understand network structure and characteristic through the relationships between nodes. A pairwise evaluation process is constructed to explore the performance of each DMU relative to another based on the DEA method. Then, the pairwise evaluation matrix that reflects the reference relationship between any two DMUs is obtained. Based on this matrix, the social network is established, and role models in the large-scale sample are identified by calculating the in-degree centrality values of all DMUs. More importantly, based on the indicator weights in DEA models, the strengths of role models are analyzed in detail. Finally, we construct the experiments with the online diagnostic data of 10599 doctors in 55 departments on the Chunyu Doctor platform in China to validate our proposed social network DEA method