Incorporating Cooperative Coevolution Algorithm with Particle Swarm Optimization in Neuro-Fuzzy System

One of the popular tools in classification area is a neuro-fuzzy model. However, there are many optimization methods used to find suitable parameters for the model. A cooperative coevolution algorithm (CCEA) is one of the popular optimization tools, since in CCEA, there are more than one species to be updated. We utilize the particle swarm optimization (PSO) to update each species. We test our algorithm on 7 UCI datasets. We found that the optimum solution from the neuro-fuzzy model with CCEA and PSO yields better accuracy than the original neuro-fuzzy model. The top 5 most important features from the model are compatible with the original neuro-fuzzy model.