Modeling and predicting the efficiency of public and private banks in Iran using an artificial neural network models, fuzzy neural networks and genetic algorithms

Continuous growth and development of the economics is considered as the main objectives which the firms are seeking to achieve. In doing so, the banks play key roles in the economic growth and development. Due to the increasing numbers of the public and private banks in Iran, predicting their efficiency has attracted significant attentions. This study aims at modeling and predicting the efficiency of the public and private banks by using artificial neural networks, Fuzzy neural networks and genetic algorithms. Using data envelopment analysis (DEA) and considering the total assets and total number of branches as the inputs of the model, the banks’ efficiency has been examined during a period from 2007 to 2011. The outputs of the model include the net profit or loss, the balance of granted credits and receivables. As the next step, the multivariate regression approach, artificial neural network, fuzzy neural network and genetic algorithms have been employed to predict the efficiency of the banks. The findings revealed that the fuzzy neural network is the most precise model in comparison with the other models of predicting efficiency. Based on the sensitivity analysis of the inputs by the neural networks, the net profit or loss has been known as the input with the highest impact on the banks’ efficiency.