Using long short-term memory networks for river flow prediction

被引:57
|
作者
Xu, Wei [1 ,2 ]
Jiang, Yanan [2 ,3 ]
Zhang, Xiaoli [4 ]
Li, Yi [1 ]
Zhang, Run [1 ]
Fu, Guangtao [2 ,5 ]
机构
[1] Chongqing Jiaotong Univ, Coll River & Ocean Engn, Natl Engn Res Ctr Inland Waterway Regulat, Chongqing, Peoples R China
[2] Univ Exeter, Coll Engn, Ctr Water Syst, Exeter EX4 4QF, Devon, England
[3] Northwest Agr & Forestry Univ, Coll Water Resources & Architectural Engn, Yanglin 712100, Shaanxi, Peoples R China
[4] North China Univ Water Resources & Elect Power, Sch Water Conservancy, Zhengzhou, Peoples R China
[5] Alan Turing Inst, 96 Euston Rd, London NW1 2DB, England
来源
HYDROLOGY RESEARCH | 2020年 / 51卷 / 06期
基金
中国国家自然科学基金; 英国工程与自然科学研究理事会;
关键词
hydrological modelling; LSTM; machine learning; river flow prediction; ARTIFICIAL NEURAL-NETWORKS; UNCERTAINTY; MODEL; LANGUAGE; SYSTEM; SWAT;
D O I
10.2166/nh.2020.026
中图分类号
TV21 [水资源调查与水利规划];
学科分类号
081501 ;
摘要
Deep learning has made significant advances in methodologies and practical applications in recent years. However, there is a lack of understanding on how the long short-term memory (LSTM) networks perform in river flow prediction. This paper assesses the performance of LSTM networks to understand the impact of network structures and parameters on river flow predictions. Two river basins with different characteristics, i.e., Hun river and Upper Yangtze river basins, are used as case studies for the 10-day average flow predictions and the daily flow predictions, respectively. The use of the fully connected layer with the activation function before the LSTM cell layer can substantially reduce learning efficiency. On the contrary, non-linear transformation following the LSTM cells is required to improve learning efficiency due to the different magnitudes of precipitation and flow. The batch size and the number of LSTM cells are sensitive parameters and should be carefully tuned to achieve a balance between learning efficiency and stability. Compared with several hydrological models, the LSTM network achieves good performance in terms of three evaluation criteria, i.e., coefficient of determination, Nash-Sutcliffe Efficiency and relative error, which demonstrates its powerful capacity in learning non-linear and complex processes in hydrological modelling.
引用
收藏
页码:1358 / 1376
页数:19
相关论文
共 50 条
  • [41] Long Short-Term Memory Neural Networks for RNA Viruses Mutations Prediction
    Mohamed, Takwa
    Sayed, Sabah
    Salah, Akram
    Houssein, Essam H.
    [J]. MATHEMATICAL PROBLEMS IN ENGINEERING, 2021, 2021
  • [42] Performance prediction in major league baseball by long short-term memory networks
    Hsuan-Cheng Sun
    Tse-Yu Lin
    Yen-Lung Tsai
    [J]. International Journal of Data Science and Analytics, 2023, 15 : 93 - 104
  • [43] Unsteady Aerodynamic Prediction for Bridges Based on Long Short-term Memory Networks
    Liu Q.-K.
    Liu S.-J.
    Zhang Z.
    Zhou X.
    Jing H.-M.
    [J]. Zhongguo Gonglu Xuebao/China Journal of Highway and Transport, 2023, 36 (08): : 56 - 65
  • [44] Prediction of Sea Ice Motion With Convolutional Long Short-Term Memory Networks
    Petrou, Zisis I.
    Tian, Yingli
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2019, 57 (09): : 6865 - 6876
  • [45] Long Short-Term Memory Networks with Multiple Variables for Stock Market Prediction
    Fei Gao
    Jiangshe Zhang
    Chunxia Zhang
    Shuang Xu
    Cong Ma
    [J]. Neural Processing Letters, 2023, 55 : 4211 - 4229
  • [46] Applying Long Short-Term Memory Networks for natural gas demand prediction
    Anagnostis, Athanasios
    Papageorgiou, Elpiniki
    Dafopoulos, Vasileios
    Bochtis, Dionysios
    [J]. 2019 10TH INTERNATIONAL CONFERENCE ON INFORMATION, INTELLIGENCE, SYSTEMS AND APPLICATIONS (IISA), 2019, : 14 - 20
  • [47] A short-term prediction model of global ionospheric VTEC based on the combination of long short-term memory and convolutional long short-term memory
    Peng Chen
    Rong Wang
    Yibin Yao
    Hao Chen
    Zhihao Wang
    Zhiyuan An
    [J]. Journal of Geodesy, 2023, 97
  • [48] A short-term prediction model of global ionospheric VTEC based on the combination of long short-term memory and convolutional long short-term memory
    Chen, Peng
    Wang, Rong
    Yao, Yibin
    Chen, Hao
    Wang, Zhihao
    An, Zhiyuan
    [J]. JOURNAL OF GEODESY, 2023, 97 (05)
  • [49] Rainfall prediction system for Bangladesh using long short-term memory
    Billah, Mustain
    Adnan, Md Nasim
    Akhond, Mostafijur Rahman
    Ema, Romana Rahman
    Hossain, Md Alam
    Galib, Syed Md
    [J]. OPEN COMPUTER SCIENCE, 2022, 12 (01) : 323 - 331
  • [50] STOCK MARKET PREDICTION USING LONG SHORT-TERM MEMORY (LSTM)
    Abu Nadif, Mohammad
    Samin, Towhidur Rahman
    Islam, Tohedul
    [J]. 2022 SECOND INTERNATIONAL CONFERENCE ON ADVANCES IN ELECTRICAL, COMPUTING, COMMUNICATION AND SUSTAINABLE TECHNOLOGIES (ICAECT), 2022,