Forecast of ship loading and unloading volume

With the development of global trade, logistics and trade information is constantly increasing. Port data such as ship characteristics, cargo types, historical loading and unloading records are also accumulated in large quantities. Therefore, it is of great significance to accurately predict the port loading volume for port management and cargo transportation. We use various characteristic project, through the establishment of the parallel fusion model of XGBoost and LightGBM, the prediction of the future ship loading volume is realized. We used the 10% data of the training set as the validation set, the RMSE score was 19,388.9,965, and the final RMSE (root-mean-square error) was 18,750.04,88The first-place score was 13,608.9,715 on the test set, ranking fifth in the total list.