We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Provide an Optimal Model for Determining and Ranking Inefficiency Factors in the Banking Industry by Combining Data Envelopment Analysis and Neural Network.
- Authors
Khojin, Gholamreza Panahandeh; Eshlaghy, Abbass Toloie; Kazemi, Mohammad Ali Afshar
- Abstract
Purpose: Data envelopment analysis is a method for comparing and evaluating the relative efficiency of decision-making units, each of which has several similar inputs and outputs. Every organization has an urgent need for an efficiency evaluation system in order to know the level of desirability and quality of its activities, especially in complex and dynamic environments. The purpose of this study is to combine two methods of data envelopment analysis and neural network in order to provide an optimal model for ranking inefficiency factors in the banking industry of Iran. Methodology: The current research is a cross-sectional descriptive study that was conducted on 32 bank branch managers. First, through the study of theoretical foundations and interviews with banking experts, performance evaluation indicators in the banking industry were identified and finalized. Further, in order to evaluate the efficiency of the units in the studied statistical population, the technique of data envelopment analysis was used, especially the modified goal programming model of data envelopment analysis, and out of the 32 bank managements studied, 3 effective managements and 29 ineffective managements were recognized. Then, branches under inefficient managements were evaluated and using the information of inefficient branches, a neural network matrix was prepared to detect inefficiency factors and the results were analyzed with different neural network models. The model with the lowest mean squared error was selected as the optimal model in order to determine the inefficiency factors. Findings: The results of combining data envelopment analysis technique, especially the modified goal programming model of data envelopment analysis with neural networks showed that self-organizing mapping neural network with hyperbolic tangent transfer function and 0.9 momentum training rule has the best performance compared to other methods in identification of inefficiency factors. After analyzing the sensitivity of the mentioned method, the indicators of the provinces' liquidity share, personnel distribution and operational costs were selected as the most important factors of inefficiency. Originality/Value: The integration of data envelopment analysis models with neural networks to evaluate the efficiency of decision-making units along with the identification of inefficiency factors can help managers of organizations in improving the inefficiency factors of decision-making units.
- Subjects
DATA envelopment analysis; BANKING industry; NEURAL circuitry; BENCHMARKING (Management); TOTAL quality management
- Publication
Journal of Decisions & Operations Research, 2022, Vol 7, Issue 4, p610
- ISSN
2538-5097
- Publication type
Article