RESEARCH ON DECK OFFICER BEHAVIOR WARNING MODEL AND ACCURACY EVALUATED BY K-NEAREST NEIGHBORS ALGORITHM

Abstract

Scientists have been spending a lot of endeavor to find useful solutions to reduce maritime accidents. Understanding and solving the causes of maritime incidents seems to be the most optimal solution to prevent maritime accidents. Maritime accident statistics show that over 80% cause of maritime accidents originate from human errors. Therefore, overcoming human errors can largely reduce the risks of maritime accidents, and is also an urgent solution. In this article, the author proposes a method to simulate warning behavior of navigator and evaluate the model accuracy using KNN (K-nearest neighbor) algorithms.