Optimal Neuro-Fuzzy model configuration
Zsolt János Viharos, Krisztián Balázs Kis · 2016
The paper is aimed to present how Neuro-Fuzzy Systems can be applied for identifying a general system model of a given problem defined by a set of variables. Neuro-Fuzzy Systems are favored in many application fields because they provide fair accuracy and their inner computational model can be interpreted through the fuzzy rules they encapsulate. The proposed input-output search algorithm is able to find optimal system configuration of an arbitrary set of variables. By placing the algorithm on a Neuro-Fuzzy basis the resulted system model became more interpretable through the inner rules of the Neuro-Fuzzy model. This makes the algorithm more interpretable by revealing more information about the inner connections between the variables of a specific problem.