Port Container Throughput Forecasting Based On Hybrid Model Of EMD-ESN-AdaBoost.TSF

Accurate port container throughput forecasting is an important means to improve port management level,port facility utilization and economic development quality.Therefore,this paper selects the port container throughput data of Shanghai Port from October 2010 to December 2019 for predictive analysis,and proposes a decomposition and integration model EMD-ESN-AdaBoost.TSF.The model first decomposes the original data to obtain 7 components and residuals.Since the first component contains too much noise,it does not make much sense for prediction.Therefore,in the prediction process,the first component and the residual are eliminated.Next,the remaining data is predicted and analyzed using the neural network of the echo state.Finally,in order to eliminate the instability of the prediction of a single neural network,this article uses an integrated method to integrate the prediction results.Empirical research shows that the combined decomposition model EMD-ESN-AdaBoost.TSF has better prediction accuracy and stability than other prediction models,and various error evaluation indicators show that the main model in this paper has the best prediction effect.The MAPE,MAE,and RMSE are the minimum values of these models.Therefore,the prediction method proposed in this paper can provide a reference for port transportation and management in practical work.