Ship Pollution Emission Prediction Based on GRU-Attention Neural Network

With the increasing global emphasis on green shipping and low-carbon energy systems, accurate prediction of maritime pollutant emissions has become critical to sustainable transport and environmental compliance. This paper proposes a GRU-Attention-based deep learning model for multivariate ship emission forecasting. By integrating AIS trajectory data, static vessel attributes, and meteorological conditions, the model leverages gated recurrent units (GRU) and an attention mechanism to capture complex temporal dependencies in emission patterns. Experimental results show that the proposed model significantly outperforms baseline LSTM and Bi-LSTM models in terms of accuracy and stability, especially during high-emission intervals. The approach contributes to intelligent maritime emission monitoring and offers a scalable framework for energy-efficient vessel operation and emission reduction strategies in new energy systems.