插值(计算机图形学)
均方误差
计算机科学
气象学
均方预测误差
算法
遥感
人工智能
数学
统计
地理
运动(物理)
作者
Guoqing Zhou,Zhou Tian
标识
DOI:10.1109/icsp58490.2023.10248859
摘要
Accurate tidal data are important for coastal city planning and maritime vessel navigation safety, but traditional tidal reconciliation and analysis methods require a large amount of data, while the more commonly used interpolation methods can only analyse the existing data and cannot predict future data. In order to solve the above problems, a tidal prediction model based on long short-term memory network (LSTM) is proposed. The tidal data from Weizhou Island, China and Clearwater Beach, USA were selected for experimental validation, and the R 2 at Weizhou Island reached 0.9741, while the RMSE was 0.14m, with a maximum prediction error of 0.27m. The R 2 at Clearwater Beach reached 0.9596, while the RMSE reached 0.05m, with a maximum error of 0.31m. Therefore, the model can effectively improve the accuracy of tidal prediction and provide a solution for tidal prediction in different sea areas.
科研通智能强力驱动
Strongly Powered by AbleSci AI