Identifying Risk-Prone Behavior of Seafarers by Using Explainable AI

Abstract

Despite the advancements in technologies for maritime navigation, maritime acci-dents are still a big problem in the industry. Extensive research has been done tostudy these accidents and try to find solutions to avoid them, but no research hastried to apply Explainable AI approaches to navigational data.

The goal of this thesis is to produce a predictive deep learning model to studynavigational data and use attention mechanisms to identify seafarer behaviors whichcould lead to accidents during the ship’s navigational operations.