Ship Sailing State Prediction Based On Stacking Model Fusion

With the expansion of international trade,the volume of maritime traffic is growing rapidly.In the sea areas with large traffic flow and complicated situation,maritime traffic accidents that cause casualties and economic losses are easy to occur.Based on the AIS data provided by the ship traffic service system,improving the safety decision-making level in ship navigation can effectively monitor and evaluate the ship navigation in real time,discover the abnormal behavior of the ship in time,and reduce the risk of water traffic accidents.Based on AIS data samples,applying data mining and machine learning theory,this paper studies the algorithm model that suitable for ship navigation state prediction,constructs the ship navigation state prediction model based on Stacking model fusion.We use the historical data of ship navigation to train the model,predict the ship's navigation in the future,and analyze the experimental results.The main research work of this paper is as follows:1.We briefly introduce the research background and current status of ship navigation safety assessment based on AIS data,and the main contents of this paper.Meanwhile,we systematically discuss the data structure of AIS and the method of data preprocessing,as well as the principle of the Stacking fusion algorithm and base classifier.2.We analyze and preprocess the AIS data to prepare the experimental data for the AIS data-based ship navigation state prediction model.And we build a ship navigation state prediction model based on Stacking model fusion,introduce the selection and experiment process of classifier in the double layer model,and optimize the model parameters.3.We compare the Stacking model with the learning curve of thebase classifier,analyze the experimental results,and calculate the prediction accuracy of the model to evaluate the prediction effect of the model.Finally,based on the model established in this paper,we use the new AIS data to analysis the experiment and result.The results show that the ship navigation state prediction model based on the Stacking model of machine learning can obtain the better prediction effect,and the prediction accuracy of the model can reach more than 97%.