The Severity of Maritime Traffic Accidents Prediction Method Based on Bilayer GA-SVM

Abstract

With the further development of maritime transportation, maritime traffic safety and the consequences of accidents have also been paid more and more attention. In order to improve the predictive accuracy of ranking ship traffic accident and solve the problem of insufficient accuracy in the existing ship accident prediction, this paper studies the idea of multi-layer classification. Based on ship accident data, a two-layer GA-SVM model is designed to achieve accurate classification of ship accident types and levels, also, comparing with other methods. The result shows a good predictive effect, which can provide a decision-making guidance for the safe navigation of ships.