Результаты исследований: Публикации в книгах, отчётах, сборниках, трудах конференций › статья в сборнике материалов конференции › научная › Рецензирование
Time Series Prediction by Reservoir Neural Networks. / Tarkov, Mikhail S.; Chernov, Ivan A.
Advances in Neural Computation, Machine Learning, and Cognitive Research IV - Selected Papers from the 22nd International Conference on Neuroinformatics, 2020. ред. / Boris Kryzhanovsky; Witali Dunin-Barkowski; Vladimir Redko; Yury Tiumentsev. Springer Science and Business Media Deutschland GmbH, 2021. стр. 303-308 (Studies in Computational Intelligence; Том 925 SCI).Результаты исследований: Публикации в книгах, отчётах, сборниках, трудах конференций › статья в сборнике материалов конференции › научная › Рецензирование
}
TY - GEN
T1 - Time Series Prediction by Reservoir Neural Networks
AU - Tarkov, Mikhail S.
AU - Chernov, Ivan A.
PY - 2021
Y1 - 2021
N2 - The tasks of forecasting time series arise in many areas of computer science. Algorithms based on machine learning do a good job of this task. In this work, we performed a comparative analysis of a number of algorithms for predicting time series by reservoir neural networks (echo-state networks) according to the forecast accuracy and the time it takes to build the forecast. To test forecasting algorithms, data sets obtained from the Mackey-Glass equation were used. The experiments showed that the sigmoidal and radial networks with a SOM projector give the most accurate forecast, but they are also the least fast. A new reservoir optimization algorithm is proposed - a direct version of the Infomax method. The functionality of the mutual information of the input and output of the reservoir is maximized. This algorithm requires non-negativity of data values, but it works much faster than the well-known iterative version of Infomax and a radial network with a SOM projector, although it slightly reduces the forecast accuracy.
AB - The tasks of forecasting time series arise in many areas of computer science. Algorithms based on machine learning do a good job of this task. In this work, we performed a comparative analysis of a number of algorithms for predicting time series by reservoir neural networks (echo-state networks) according to the forecast accuracy and the time it takes to build the forecast. To test forecasting algorithms, data sets obtained from the Mackey-Glass equation were used. The experiments showed that the sigmoidal and radial networks with a SOM projector give the most accurate forecast, but they are also the least fast. A new reservoir optimization algorithm is proposed - a direct version of the Infomax method. The functionality of the mutual information of the input and output of the reservoir is maximized. This algorithm requires non-negativity of data values, but it works much faster than the well-known iterative version of Infomax and a radial network with a SOM projector, although it slightly reduces the forecast accuracy.
KW - Forecast
KW - Maximization of mutual information
KW - Reservoir neural networks
KW - Time series
UR - http://www.scopus.com/inward/record.url?scp=85093092403&partnerID=8YFLogxK
UR - https://www.mendeley.com/catalogue/42b48358-a880-3f68-9011-c04082457209/
U2 - 10.1007/978-3-030-60577-3_36
DO - 10.1007/978-3-030-60577-3_36
M3 - Conference contribution
AN - SCOPUS:85093092403
SN - 9783030605766
T3 - Studies in Computational Intelligence
SP - 303
EP - 308
BT - Advances in Neural Computation, Machine Learning, and Cognitive Research IV - Selected Papers from the 22nd International Conference on Neuroinformatics, 2020
A2 - Kryzhanovsky, Boris
A2 - Dunin-Barkowski, Witali
A2 - Redko, Vladimir
A2 - Tiumentsev, Yury
PB - Springer Science and Business Media Deutschland GmbH
T2 - 22nd International Conference on Neuroinformatics, 2020
Y2 - 12 October 2020 through 16 October 2020
ER -
ID: 25677048