Djumanov, O. I., Kholmonov, S. M., & Ganiev, J. M.
Optimization of data processing based on neural networks and properties of non-stationary objects. |
 |
Full Article: PDF
Scientific Object Identifier: http://s-o-i.org/1.1/TAS-04-96-35
DOI: https://dx.doi.org/10.15863/TAS.2021.04.96.35
Language: English
Citation: Djumanov, O. I., Kholmonov, S. M., & Ganiev, J. M. (2021). Optimization of data processing based on neural networks and properties of non-stationary objects. ISJ Theoretical & Applied Science, 04 (96), 165-168. Soi: http://s-o-i.org/1.1/TAS-04-96-35 Doi: https://dx.doi.org/10.15863/TAS.2021.04.96.35 |
Pages: 165-168
Published: 30.04.2021
Abstract: Methodological foundations have been developed for the creation of methods for optimizing the analysis of data of non-stationary objects based on neural networks (NN) and genetic algorithms (GA) with mechanisms of simplified search, adjusting the weights of neurons, coefficients of synaptic connections, activation functions, the number of neurons in the layers of NN. Improvement and development of methods for optimizing learning of neural networks based on the synthesis of GA s, the use of operators of crossworing, mutation, inversion, selection of the best individuals, generation of the initial and subsequent populations has been carried out. Mechanisms for generating a bank of individuals for the formation of rational sets of parameters and training samples based on a knowledge base with fuzzy inference rules are proposed.
Key words: data processing, neural network, genetic algorithm, tuning, optimization, bank of individuals, database, knowledge base.
|
