An Online Self-constructing Fuzzy Neural Network with Restrictive Growth

摘要

In this paper, a novel paradigm, termed online self constructing fuzzy neural network with restrictive growth (OSFNNRG) which incorporates a pruning strategy into new growth criteria, is proposed. The proposed growing procedure without pruning not only speeds up the online learning process but also results in a more parsimonious fuzzy neural network while comparable performance and accuracy can be achieved by virtue of the growing and pruning mechanism. The OSFNNRG starts with no hidden neurons and parsimoniously generates new hidden units according to the proposed growth criteria as learning proceeds. In the parameter learning phase, all the free parameters of hidden units, regardless of whether they are newly created or originally existing, are updated by the extended Kalman filter (EKF) method. The performance of the OSFNNRG algorithm is compared with other popular approaches like OLS, RBF-AFS, DFNN and GDFNN in nonlinear dynamic system identification. Simulation results demonstrate that the learning speed of the proposed OSFNNRG algorithm is faster and the network structure is more compact with comparable generalization performance and accuracy.