摘要
作为重要的海洋环境参数,有效波高(SWH)的精准预测在海洋工程中具有重要意义。针对深度学习易陷入局部最优等问题,提出了一种融合深度-宽度的SWH混合预测方法。利用集成经验模态分解对SWH进行预处理,改善深度学习预测中普遍存在的滞后性问题,同时将宽度学习系统与深度学习中的长短时记忆网络相融合,旨在提高预测精度。实验结果表明:与现有方法相比,基于本文模型的SWH预测方法不但在均方根误差、平均绝对误差等评价指标上得到有效提升,且具有较好的鲁棒性。
As an important oceanic environmental parameter,the accurate prediction of significant wave height(SWH)is of great significance in marine engineering.To address the issues such as deep learning easily falling into local optima,this paper proposes a hybrid swH prediction method that combines depth and width.The method utilizes ensemble empirical mode decomposition to preprocess SWH,aiming to improve the common lag problem in deep learning prediction.Furthermore,it integrates the broad learning system with the long short-term memory network in deep learning to enhance prediction accuracy.The experimental results show that the hybrid SwH prediction method not only effectively improves the evaluation indicators such as root mean square error and mean absolute error,but also exhibits good robustness compared with existing methods.
作者
闫加宁
唐家昕
任硕
董昌明
韩莹
YAN Jianing;TANG Jiaxin;REN Shuo;DONG Changming;HAN Ying(School of Automation,Nanjing University of Information Science and Technology,Nanjing,210044,China;Southern Marine Science and Engineering Guangdong Laboratory(Zhuhai),Zhuhai,519000,China;School of Ocean Science,Nanjing University of Information Science and Technology,Nanjing,210044,China;School of Electronics and Electrical Engineering,Wuhan Textile University,Wuhan 430200,China)
出处
《海洋测绘》
CSCD
北大核心
2024年第2期41-45,共5页
Hydrographic Surveying and Charting
基金
国家自然科学基金(62076136)
南方海洋科学与工程广东省实验室(珠海)基金(SML2020SP007)。
关键词
有效波高预测
深度学习
宽度学习系统
集成经验模态分解
长短时记忆网络
prediction of significant wave height
deep learning
broad learning system
ensemble empirical mode decomposition
long short-term memory network
