A Short-term Forecasting Model of Elevated Duct Based on the Long Short-Term Memory Network

In recent years, the long-range no-relay maritime transmission of wireless signals based on atmospheric ducts has become a prominent research topic. To improve the forecasting accuracy of atmospheric duct characteristics supporting the above transmission way, we proposed a short-term forecasting model for elevated duct strength based on the Long Short-Term Memory (LSTM) Network. This model has demonstrated significant advantages in time series analysis, as it can capture long-term dependencies within the data. This model leverages the SPARC dataset to train and test an LSTM model for forecasting the future strength trends of elevated ducts. The model’s performance is evaluated using Root-Mean Square Error (RMSE) and Mean Absolute Error (MAE), with results showing RMSE and MAE values of 4.90 M -unit and 3.30 M -unit, respectively.