Open AccessDevelopment of Stability Control Mechanisms in Neural Network Forecasting Systems
Author(s) -
I. V. Miloserdov,
Dmitriy Miloserdov
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1864/1/012105
Subject(s) - stability (learning theory) , artificial neural network , computer science , balance (ability) , control (management) , artificial intelligence , order (exchange) , machine learning , neuroscience , psychology , economics , finance
The problem of ensuring the stable functioning of time series forecasting systems based on streaming recurrent neural networks with controlled elements is considered. The mechanisms necessary and sufficient for its maintenance are derived, which involve maintaining the balance of the learning history and modifying the synapse learning rules in order to establish a balance between positive and negative potential. The results of experiments to assess the accuracy of forecasting are presented.