Tidal Forecasting Based on ARIMA-LSTM Neural Network

In order to improve the accuracy of tidal forecasting, a neural network deep learning method is proposed, which combines LSTM with ARIMA to predict the tidal height. ARIMA-LSTM model also uses the prediction value of harmonic analysis as the factor, and then considers the wind speed and sea water temperature to predict. Firstly, the strong nonlinear relationship ability of LSTM is used to predict, and then the data is linearly fitted and predicted through ARIMA model, so this step can be regarded as the linearization process of error. The experimental results verify the feasibility and effectiveness of the method, and obtain good simulation results, which verify that the prediction accuracy of the model is high and that of the traditional method.