Iranian Remote Sensing and GIS Society / Shahid Beheshti University Iranian Journal of Remote Sensing and GIS 2008-5966 13 4 2022 02 20 Finding Optimal Contextual Parameters for Real-Time Vessel Position Prediction Using Deep Learning Finding Optimal Contextual Parameters for Real-Time Vessel Position Prediction Using Deep Learning 89 100 101688 10.52547/gisj.13.4.89 FA Ali Asghar Alesheikh Full Prof., Dep. of Geomatics Engineering, K.N. Toosi University of Technology, Tehran Saeed Mehri K. N. Toosi University of Technology Journal Article 2019 09 20 About 80% of world transportation happens at sea. Therefore the safety of vessels, in particular<br />during vessels’ movement, is crucially important. As different contextual parameters affect vessels’<br />movement, selecting optimal contextual parameters is one of the main changes in vessels’ Context-<br />Aware movement analysis. Toward this end, a Long Short-Term Memory (LSTM) network is used<br />for wrapper feature selection to identify optimal contextual parameters for vessels’ movement<br />prediction. To do this, the Automatic Identification System (AIS) dataset from the eastern coast of the<br />United States of America collected from December 2017 is used. All possible combinations of three<br />contextual parameters, including speed, course and vessels’ presence probability in different positions<br />at sea, were evaluated using the wrapper method in the LSTM network. In all evaluations, 70% of<br />data was used for training and the remaining for cross-validation. The results selected speed and<br />presence probability as optimal contextual parameters for vessel movement prediction. The model<br />trained with optimal contextual parameters is 26.98% more accurate than a model trained with all<br />available contextual parameters and 16.14% better than a model without contextual parameters.<br />Therefore, selecting optimal parameters from available contextual parameters can help improve the<br />accuracy of vessels’ predictions. Keywords: Context-Aware, Long Short-Term Memory, Automatic<br />Identification System, wrapper, Movement prediction, Context. About 80% of world transportation happens at sea. Therefore the safety of vessels, in particular<br />during vessels’ movement, is crucially important. As different contextual parameters affect vessels’<br />movement, selecting optimal contextual parameters is one of the main changes in vessels’ Context-<br />Aware movement analysis. Toward this end, a Long Short-Term Memory (LSTM) network is used<br />for wrapper feature selection to identify optimal contextual parameters for vessels’ movement<br />prediction. To do this, the Automatic Identification System (AIS) dataset from the eastern coast of the<br />United States of America collected from December 2017 is used. All possible combinations of three<br />contextual parameters, including speed, course and vessels’ presence probability in different positions<br />at sea, were evaluated using the wrapper method in the LSTM network. In all evaluations, 70% of<br />data was used for training and the remaining for cross-validation. The results selected speed and<br />presence probability as optimal contextual parameters for vessel movement prediction. The model<br />trained with optimal contextual parameters is 26.98% more accurate than a model trained with all<br />available contextual parameters and 16.14% better than a model without contextual parameters.<br />Therefore, selecting optimal parameters from available contextual parameters can help improve the<br />accuracy of vessels’ predictions. Keywords: Context-Aware, Long Short-Term Memory, Automatic<br />Identification System, wrapper, Movement prediction, Context. https://gisj.sbu.ac.ir/article_101688_99a1b8cea5f967844a24000153bf4347.pdf