Adaptive Processing of Technological Time Series for Forecasting Based on Neuro-Fuzzy Networks

Ibragimovich, Jumanov Isroil
Baxromovna, Melieva Mokhinur

IJHCS
Vol. 4 No. 2 (2022): IJHCS

Methodological bases for identification, data processing for forecasting technological time series based on the synthesis of soft computing apparatus (dynamic models, neural networks, neuro-fuzzy networks, genetic algorithms) in various combinations have been developed. A generalized prediction optimization algorithm based on a hybrid model with mechanisms for determining and adjusting the weights of neurons, coefficients of synaptic connections, activation functions, determining the number of layers and neurons in the layers of neural networks with a rational architecture is proposed.