Social network analysis for cross‐evaluation in data envelopment analysis. Issue 9 (9th June 2022)

• Record Type: Journal Article
• Authors: Ang, Sheng; Wu, Hui; Chen, Menghan; Yang, Feng
• Other Names: Herrero Álvaro guestEditor.; Urda Daniel guestEditor.; Sedano Javier guestEditor.; Quintián Héctor guestEditor.; Corchado Emilio guestEditor.; Ahmed Syed Hassan guestEditor.; Khan Murad guestEditor.; Guibene Wael guestEditor.
• Is Part Of: Expert systems. Volume 39:Issue 9(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-06-09
• Subjects: cross‐efficiency evaluation -- data envelopment analysis -- ranking -- social network analysis; Expert systems (Computer science); 006.33
• DOI: 10.1111/exsy.13063
• Languages: English
• ISSNs: 0266-4720

Abstract

Abstract: As one of the important extensions of data envelopment analysis (DEA) methodology, cross‐evaluation combines self‐evaluation and peer‐evaluation processes for each decision making unit (DMU) to effectively eliminate unrealistic weight schemes and differentiate all the units for ranking. Differing from previous research, which mainly pays attention to either secondary goals for generating weights for peer evaluation or aggregation of self and peer‐evaluation efficiencies, our study focuses on evaluating the relationships among all DMUs. If we view each unit as a node to be evaluated, then peer‐evaluation among pairs of DMUs forms a bidirectional network among all the DMUs. Based on this idea, we introduce a data‐driven tool combining social network analysis and cross‐evaluation to examine self‐evaluated and peer‐evaluated scores in the network. Using the hyperlink‐induced topic search algorithm, we develop the concepts of self‐evaluation and peer‐evaluation objectivity coefficients based on how well DMUs accept the analysis results. The rankings based on our self‐evaluation and peer‐evaluation scores adjusted with objectivity coefficients are shown to be consistent. Our study provides a novel perspective for cross‐evaluation and DEA research. An application case study illustrates our new method.