Specialized forecasting and diagnostic neuro-fuzzy network is considered, which is a generalization of a multilayer perceptron, recurrent Williams-Zipser network and forecasting multilayer neuro-fuzzy network. The corresponding learning and diagnostic algorithm is proposed. Such hybrid approach allows simultaneously to decide the tasks of short-term prognostication and control of consumption of electric energy.
neuro-fuzzy network, prognostication of temporal rows, technical diagnostics
“Prognoziruiushche-diagnostiruiushchaia rekurrentnaia neiro-fazzi set v zadache kontrolia potrebleniia elektroenergii”,
Information Processing Systems,