Determinants of Islamic Banks' Profitability Using Panel Data Analysis and ANFIS Approaches in Saudi Arabia

The purpose of this study is to determine the relationship between profitability and financial ratios of Islamic banks in Saudi Arabia. To accomplish this goal, quarterly data of four Islamic banks from 2009 to 2017 were considered. The Artificial Neural Network (ANN) and Fuzzy System, or ANFIS (Adaptive Neuro- Fuzzy Inference System), was employed to predict the profitability of financial ratios. To determine impacts of financial ratios on the profitability measures, the study includes panel data analysis. The average prediction error for the ANFIS model of return on asset (y1) with four rules was 0.3537%, and the return on equity (y2) with five rules was 0.31829%. The results showed that all the explanatory variables, except stock capital gain ratio, have a significant positive relation with profitability measure of either return on asset or return on equity of Islamic banks in Saudi Arabia. However, only total equity to total asset and earning per share ratios have relation with both the profitability measures. The results of descriptive statistics, multiple regression, and ANFIS models established that successful outcome can be obtained for y1 and y2. Therefore, this study will be beneficial not only for the literature, but also for the investors and executives of Islamic banking.