Shenzhen Vessel Traffic Flow Prediction Based On Least Squares Support Vector Machines

China has a vast sea area,there are many ships and crew.With the rapid development of China’s waterway transportation and the improvement of the degree of development and utilization of the sea,the increase in the number of ships in distress and the number of people in distress are increasing year by year.Navigation conditions have also become increasingly complex,and the maritime security situation has become more severe.It is in this context that this paper studies the ship traffic flow forecasting method,and a more accurate and efficient forecasting model can ensure the timeliness of water safety supervision and improve the safety of water safety supervision.Accurate and timely prediction of water traffic flow can effectively reduce the incidence of ship accidents and ensure the safety of life and property.At the same time,it provides support and basis for policy formulation of maritime administrative departments and decision-making planning of port enterprises.On the basis of summarizing the existing research on the vessel traffic flow forecast,this paper expounds the ship traffic flow forecast.Firstly,in order to improve the scientificity and accuracy of the input variables of the prediction model,reduce the influence of subjective factors,the gray correlation analysis is used to screen the influencing factors of the traffic flow of the ship,and more reasonable factors are selected as the input variables of the prediction model.Then,to avoid the shortcomings and limitations of the existing ship traffic flow prediction model,the least squares support vector machine is selected as the prediction method,and the particle swarm optimization algorithm is used as the optimization algorithm to construct the PSO-LSSVM ship traffic flow prediction model,and the monthly ship traffic flow in Shenzhen is used as empirical analysis.The result compared with the GA-LSSVM predictive model and the GA-WNN predictive model.The results show that the prediction error of the PSO-LSSVM prediction model is lower than that of the GA-LSSVM prediction model and the GA-WNN prediction model,and the error is reduced by 9.98%and 88.02%,respectively,indicating that the LSSVM is easier to obtain accurate than the neural network.The LSSVM optimized by the particle swarm optimization algorithm tends to be globally optimal,and the PSO-LSSVM prediction model can predict the traffic flow of ships in specific sea areas.