Integration of neural network algorithms, nonlinear dynamics models, and fuzzy logic methods in prediction problems
P. A. Borisov, Г.П. Виноградов, Н.А. Семенов · Journal of Computer and Systems Sciences International · 2008
The problem of designing the architecture of a fuzzy hybrid system that allows one to solve the online prediction problem in conditions of strongly grouped and irregularly distributed data is considered. It is shown that parallel updating of premises and conclusions in the knowledge base hastens learning and improves the operating quality of the system.