Data envelopment analysis with fuzzy inference system to evaluate the performance of higher education institutions

Abstract

Evaluating the performance of higher education institutions holds a significant position due to their critical responsibilities. This study aims to design an integrated model utilizing Data Envelopment Analysis (DEA) and a fuzzy inference system to comprehensively assess higher education institutions by considering various input and output variables from multiple perspectives. This approach addresses the challenges of evaluating decision-making units efficiently when faced with numerous inputs and outputs. Various methods, including reducing input and output variables, have been proposed in the literature to improve this issue, typically relying on either objective or subjective approaches. To enhance discriminative power, factor analysis was employed to consider data patterns, and the fuzzy inference system was used to incorporate expert judgments. Consequently, the proposed model provides more realistic weights for inputs and outputs by integrating these two approaches. Moreover, the model’s ability to simultaneously use qualitative and quantitative variables as inputs and outputs of the DEA is another significant advantage. The proposed model was tested on higher education institutions as a case study. The results indicate that the discriminative power of decision-making units is higher compared to other variable reduction models, including subjective approaches such as DEA-AHP and DEA-MEAN, objective approaches like DEA-PCA, and combined approaches like DEA-WPCA. Furthermore, the Kruskal-Walli’s test results confirm that the proposed model yields a lower average efficiency ranking compared to other models, validating its superiority and originality.