大连港口货物吞吐量的时间序列分析

his article gives the knowledge of Box-Jenkins time series analysis model, and makes a forecasting for Dalian port cargo throughput. Specific process: First of all, makes a judgement of the time series trend in accordance with the original data, and uses the improved combination model to remove the index trends and cyclical trends, and achieves its estimated value parameters using nonlinear least squares. Then sets order of ARMA(p,q) model for the random part, estimates parameters sand tests the model, and determines the time series model of Dalian port cargo throughput. Finally, forecasts for short-term cargo throughput making use of the model and gives the conclusions of the quantitative and qualitative. Model analysis, modeling and forecasting is based on the Matlab toolbox and procedures designed to achieve. Theoretical analysis and application of this article show that the model established by the model and the improvement are feasible, and the results achieve the effect expected.