宁波港集装箱吞吐量预测模型的选择

The Selection of Forecast Model of Ningbo Port’s Container Throughput

宁波港集装箱运输业务的发展状况对宁波经济和浙江经济有重要影响,所以如何准确预测宁波港集装箱吞吐量,就成为一个重要课题。之前许多学者运用关联因素预测法、GM(1,1)灰度模型、三次指数平滑法等方法进行预测,但预测精度不高。文章运用时间序列预测模型中的ARIMA模型,先后进行49次建模尝试,最终建立最优ARI MA(4,2,4)模型,并顺利通过一系列检验。运用该模型,对宁波港集装箱吞吐量进行预测,精度较高。同时预测结果显示,近两年宁波港集装箱吞吐量将继续保持低速增长。

The development of port＇s container transportation business has important implications on economy of Ningbo and Zhejiang, so how to accurately predict the container throughput of Ningbo port becomes an important subject. Many scholars predict by correlation factors predicting method, GM （1, 1） gray model and three exponential smoothing method, but the prediction precision is not high. Applying ARIMA model, this paper has carried out 49 modeling, eventually establishes the optimal ARIMA （4, 2, 4） model, and successfully passes a series of tests. Using this model, Ningbo port container throughput is predicted, and the precision is good. At the same time, the predicted results show that in nearly two years, Ningbo port＇s container throughput will continue to maintain a low growth.