深度学习神经网络及其在海洋环境信息挖掘预测中的应用  

Deep learning neural network and its application in marine environmental information mining prediction

作　　者：徐凌宇[1] 张高唯 江湾湾 李新绢 王蕾[2] 

机构地区：[1]上海大学计算机学院,上海200444 [2]国家海洋局东海信息中心,上海200137

出　　处：《海洋信息》2018年第1期17-23,35,共8页Marine Information

基　　金：中国工程院知识中心项目-海洋工程专业知识服务系统

摘　　要：文章介绍了深度学习在时间序列分析挖掘中的应用、深度学习的概况与特征以及深度学习方法的最新国内外应用,特别是在海洋信息挖掘预报中的新应用;给出了典型深度学习神经网络的原理、结构与训练方法,随后给出针对海洋环境信息提出的3种深度神经网络模型;并分析了方法的先进性与可行性。This paper introduces the application of deep learning in time series analysis and mining, the general status and characteristics of deep learning, and the latest application of deep learning methods at home and abroad,especially in the field of ocean information mining and prediction. In addition, this article also introduces the principle, structure and training method of typical deep learning neural network. Then three kinds of deep neural network models for marine environmental information are given, and the advantage and feasibility of the methods are also analyzed.

关 键 词：深度学习 神经网络 海洋环境信息 挖掘 预测 

分 类 号：TP389.1[自动化与计算机技术—计算机系统结构]