Performance Modelling on Banking System: A Data Envelopment Analysis-Artificial Neural Network Approach

With changing banking environment, the efficiency of the operational function of bank is of critical importance and needs timely watch. Apart from measuring the operational performance of banks using DEA approaches, the banking sector today is more inclined to predictive analytics to identify their future performance and improve their competitiveness well in advance. In this sequel, the present paper proposes hybridisation of Data Envelopment Analysis and Artificial Neural Network Approaches for operational performance measurement and prediction for Indian banks using the five-year (2015 to 2019) dataset. Non-oriented non-radial DEA model is adopted in the present study, attempting to provide decision-makers the discretion to identify slacks in performance by maximising outputs and minimising inputs. This can identify causes of inefficiency and suggest necessary steps for improvement. In addition to DEA findings, the paper performs prediction task for obtained efficiency scores. Finding of will be advantageous for policymakers, managers of banking industry for predicting future operational performance of banks until they are able to make required changes for its improvement.