Оптимізація параметрів гібридної нейро-фаззі системи (ANFIS) для підвищення стійкості до шуму в задачах розпізнавання образів

Abstract

This paper presents an optimized Adaptive Neuro-Fuzzy Inference System (ANFIS) for improved noise resilience in image recognition. ANFIS integrates fuzzy logic and neural networks to handle uncertain data, but its performance heavily depends on parameter selection. To enhance robustness, we apply metaheuristic optimization techniques such as Genetic Algorithms (GA) and Particle Swarm Optimization (PSO). Experimental results demonstrate that the optimized ANFIS outperforms conventional models by maintaining higher classification accuracy and better resistance to noise. The proposed approach is suitable for applications in medical imaging, security, and industrial automation.