Forecasting Delivered Ship Waste to Petroleum Ports Using RNN Models

The objective of this study is to compare the performance of three deep learning algorithms (RNN: Recurrent Neural Networks, LSTM: Long-Shirt Term Memory and GRU: Gated Recurrent Unit) based on three statistical error metrics (MAE: Mean Absolute Error, RMSE: Root Mean Square Error and MAPE: Mean Absolute Percentage Error) in predicting the quantities of solid waste disposed of at a petroleum terminal based in Morocco. LSTM model was found to provide the most accurate and precise results among the deep learning techniques investigated in this study, followed by GRU, as RNN came last.