Time Series Prediction by Reservoir Neural Networks

Mikhail S. Tarkov, Ivan A. Chernov

Результат исследования: Публикации в книгах, отчётах, сборниках, трудах конференцийстатья в сборнике материалов конференциинаучнаярецензирование


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.

Язык оригиналаанглийский
Название основной публикацииAdvances in Neural Computation, Machine Learning, and Cognitive Research IV - Selected Papers from the 22nd International Conference on Neuroinformatics, 2020
РедакторыBoris Kryzhanovsky, Witali Dunin-Barkowski, Vladimir Redko, Yury Tiumentsev
ИздательSpringer Science and Business Media Deutschland GmbH
Число страниц6
ISBN (печатное издание)9783030605766
СостояниеОпубликовано - 2021
Событие22nd International Conference on Neuroinformatics, 2020 - Moscow, Российская Федерация
Продолжительность: 12 окт 202016 окт 2020

Серия публикаций

НазваниеStudies in Computational Intelligence
Том925 SCI
ISSN (печатное издание)1860-949X
ISSN (электронное издание)1860-9503


Конференция22nd International Conference on Neuroinformatics, 2020
СтранаРоссийская Федерация


Подробные сведения о темах исследования «Time Series Prediction by Reservoir Neural Networks». Вместе они формируют уникальный семантический отпечаток (fingerprint).