DESIGN AND IMPLEMENTATION OF ANFIS CONTROLLERS FOR STABILIZING FINANCIAL SYSTEMS: A COMPARATIVE STUDY WITH NONLINEAR FEEDBACK CONTROL

The study revisits the well-known Bouali chaotic financial model, which is characterized by nonlinear dynamics. As a benchmark, the nonlinear feedback control method is implemented and compared with an Adaptive Neuro-Fuzzy Inference System (ANFIS) controller. The ANFIS model is trained using 250 data samples derived from the nonlinear feedback controller and divided into training, validation, and testing subsets. The proposed ANFIS controller demonstrates superior stabilization performance by effectively eliminating chaotic behavior, ensuring stability, and achieving faster convergence than the traditional nonlinear feedback method. Quantitative results confirm this improvement: the ANFIS controller achieved very low Root Mean Square Error (RMSE) values, such as 8.78×10−5 for training and 1.37×10−4 for validation in the profit control input, highlighting its learning accuracy. Additionally, the ANFIS maintained stability even with a reduced number of controllers, demonstrating robustness and adaptability. These findings emphasize the potential of ANFIS controllers to provide efficient and reliable chaos control in complex financial systems.