Application of machine learning methods in maritime safety information classification

摘要

In order to ensure the safety of maritime navigation, International Maritime Organization (IMO) developed the Global Maritime Distress and Safety System (GMDSS), which includes an important Maritime Safety Information (MSI) Broadcasting System that can broadcast navigational warnings and other crucial information. However, in most sea areas, the broadcast messages are still not pre-classified. Seafarers need to identify the theme of those received messages artificially, with low efficiency and accuracy. To solve this problem, several machine-learning based solutions are presented, implemented, and compared. Thousands of Navigational Telex (NAVTEX) messages, from 2011 to 2017, are used in the training and evaluation progress. The classifiers are compared in terms of accuracy, efficiency, precision and recall rate, and F-Measure respectively; and one of them is chosen as the optimal classifier. The chosen classifier performs well on solving the NAVTEX classification problem, and may further be used in other similar problems.