%P 1-4
%I Institute of Electrical and Electronics Engineers Inc.
%A K. Osawa
%A H. Yamaguchi
%A M. Umair
%A M.A. Hashmani
%A K. Horio
%T Wave Height and Peak Wave Period Prediction Using Recurrent Neural Networks
%R 10.1109/ICCI51257.2020.9247805
%D 2020
%L scholars12641
%J 2020 International Conference on Computational Intelligence, ICCI 2020
%O cited By 4; Conference of 2020 International Conference on Computational Intelligence, ICCI 2020 ; Conference Date: 8 October 2020 Through 9 October 2020; Conference Code:164916
%X In this paper, we applied a recurrent neural network to predict a wave height and a peak wave period for next 24 hours from only those last 24 hours. We adopted LSTM as the network structure and used statistic gradient decent method and adaptive moment estimation method as the learning methods. It was difficult to estimate short-time fluctuations because only the wave height and period data were used as inputs, but it was shown that the wave height and peak wave period within the next 2 hours can be predicted with an accuracy within 20 percent in error. © 2020 IEEE.
%K Intelligent computing; Learning systems; Water waves, Learning methods; Moment estimation method; Network structures; Wave heights; Wave period, Long short-term memory