The paper solves the problem of selecting the recurrent neural network architecture to construct geomagnetic SME index architecture. The description is given for the architectures providing the processing of longterm dependencies in data. Neural network performance quality was estimated on the basis of errors. The Dropout layer effect on the modeling results is shown separately. The constructed neural network models can be used to forecast the SME index variations. [ABSTRACT FROM AUTHOR]

Copyright of EPJ Web of Conferences is the property of EDP Sciences and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)