Application of Multi-Layer Recurrent Neural Network and fuzzy time series in input/output prediction of DEA models: Real case study of a Commercial Bank

Data envelopment analysis (DEA) is a non-parametric technique based on linear programming in which multiple inputs and outputs are simultaneously used in order to measure technical efficiency. All of the research efforts have used DEA approach as a tool for evaluating what has been occurred up to the present time. However, due to the time lag, it is usually too late for the Decision Making Units (DMUs) under assessment to react the outcomes timely. In this paper a couple of forecasting procedures (fuzzy time series and Recurrent Neural Network (RNN)) have been proposed in order to supply oncoming measures of DMUs. The outputs of aforementioned forecasting procedures have been distinctively used through supposed DEA model to evaluate technical efficiency and identify the reference set for inefficient DMUs. The procedure has efficiently been applied in a real case study concerning Commercial Bank. The resultant outcomes are promising and impressive for aforementioned large scale and complicated real case study.