Evaluation of Radial Basis Functions in Fuzzy Interface Models for Predictive Training and Testing

In this study, we explore the application of radial basis networks within a fuzzy interface model to train and test predictive capabilities in simulation environments. The methodology involves the use of triangular membership functions to ensure precise parametric adjustments, resulting in a strong correlation between actual and predicted data values. The research highlights the limitations of radial basis networks in maintaining prediction control, particularly noting the challenges faced during deeper simulation phases. Through comprehensive simulation maps, we demonstrate the model's initial performance efficacy, followed by a noticeable decline in accuracy as simulations progress. These findings provide valuable insights into the optimization of fuzzy models for enhanced prediction accuracy.