Predicting Bank Profitability in Iran by Fuzzy Inference System

The main purpose of this study is to develop a Fuzzy inference system to predict bank profitability in Iran and help investors in their investment decisions. For this purpose, the main effective variables on bank profitability, including facilities, deposits, manpower costs, and assets were recognized. In the next step, the data of 13 banks were collected from 2001 to 2011. The membership functions and Fuzzy rules were developed in the MATLAB software and then, Fuzzy inference system was developed. The findings revealed that the system has an appropriate performance in predicting profitability of Iranian banks and rarely makes any error in this area. The predicted profitability of many banks has increased during the study period and also the predicted profitability of private banks was more than public banks. The banks of Industry and Mine and Karafarin Bank had the least profitability and Mellat Bank had the highest. Finally, Post Bank had the most errors while Mellat Bank had the fewest errors. Keywords: Fuzzy Inference System, Bank profitability, Membership Function, Linguistic labeling, Facilities JEL Classification: E59, C61, G24