Phân tích chuỗi dữ liệu nghiệm triều sử dụng mạng nơ ron hồi tiếp với nút có cổng (GRU)

This study applied the GRU (Gated Recurrent Unit) model, a type of Recurrent Neural Network (RNN), to analyze tidal data collected at the Vung Tau tide gauge station from January 1, 1999, to December 31, 2022.Experimental results show that the statistical performance of the model is very promising: RMSE = 2.2 mm; MAE = 0.5 mm; Kappa = 0.98, and F-Score = 0.96.Additionally, to verify the effectiveness of the model, the dataset was divided into training, testing, and prediction datasets.Statistical results indicate that the GRU model can reliably predict tidal values using statistical features of the prediction dataset, such as RMSE = 0.06mm, MAE = 0.05 mm.The study also highlights the necessity of implementing outlier detection and elimination for the input dataset to achieve high prediction accuracy.