Adaptive Learning Algorithm for Hybrid Fuzzy System
Proceedings of the International Conference "Traditional and Innovations in Sustainable Development of Society"
2002
Aleksandrs Vališevskis
In this paper the possibility of improving convergence time of algorithms intended for tuning
parameters of fuzzy system with inference mechanism realized with the help of adaptive network is considered. A new
algorithm is proposed, which allows to decrease the number of iterations during learning process and to substantially
decrease the number of computational operations that have to be performed during single iteration. Furthermore,
analytical data is presented and it’s shown how to reduce the computational load in the case if the proposed algorithm
is being used.
Atslēgas vārdi
adaptive learning algorithms, adaptive network, Adaptive Network Based Fuzzy Inference System, neuro-fuzzy systems, Resilient backpropagation
Vališevskis, A. Adaptive Learning Algorithm for Hybrid Fuzzy System. No: Proceedings of the International Conference "Traditional and Innovations in Sustainable Development of Society", Latvija, Rēzekne, 28. Feb-2. Mar., 2002. Rēzekne: Rēzeknes Augstskola, 2002, 281.-287.lpp.
Publikācijas valoda
English (en)