摘要
本文综合运用ARIMA预测模型和LSSVR预测模型,提出了一种集成预测模型,并将该模型应用于上海港的集装箱吞吐量预测研究中。此外,采用不同的参数估计方法估计ARIMA模型的参数,得到了两种ARIMA预测模型。研究表明,集成预测模型可以提高预测模型的准确性,不同的估计方法也会影响模型的预测表现。
By using ARIMA prediction model and LSSVR prediction model, this paper proposes an integrated prediction model,and applies it to the prediction of container throughput of Shanghai port. In addition, parameters of the ARIMA model are estimated by different parameter estimation methods, and two ARIMA prediction models are obtained. The research shows that the integrated prediction model can improve the accuracy of the prediction model, and different estimation methods will also affect the prediction performance of the model.
引文
[1]刘钰.基于VMD-ARIMA-HGWO-SVR组合模型的港口集装箱吞吐量预测[D].兰州:兰州大学(硕士学位论文),2018.
[2] Mak K L, Yang D H. Forecasting Hong Kong's Container Throughput with Approximate Least Squares Support Vector Machines[J]. World Congress on Engineering, 2007(1):7-12.
[3]刘雷丽,蒋惠园,张栓柱.基于数据预处理GM 1,!1"模型的深圳港集装箱吞吐量预测[J].水运工程,2009(2):83-86.
[4]薛俊强.宁波港集装箱吞吐量预测模型的选择[J].华东经济管理,2013,27(5):169-172.
[5]田雪,王丹丹,王锐月,等.基于灰色模型的港口吞吐量预测研究——以曹妃甸港口为例[J].数学的实践与认识,2018,48(4):280-284.
[6]朱念,陈东升,何昌勤,等.基于灰色GM 1,!N"模型的广西北部湾港口物流预测研究[J].数学的实践与认识,2017,47(23):303-310.
[7] Wen-Yi Peng, Ching-Wu Chu. A comparison of univariate methods for forecasting container throughput volumes[J]. Mathematical and Computer Modelling, 2009,50(7):1045-1057.
[8]鲁渤,杨显飞,汪寿阳.基于情境变动的港口吞吐量预测模型[J].管理评论,2018,30(1):195-201.
[9] Gang Xie, Shouyang Wang, Yingxue Zhao, et al. Hybrid approaches based on LSSVR for container throughput forecasting:A comparative study[J]. Applied Soft Computing Journal, 2013,13(5):2232-2241.
[10]施泽军,李凯.基于灰色模型和指数平滑法的集装箱吞吐量预测[J].重庆交通大学学报(自然科学版),2008(2):302-304,332.
[11]赵尚威,周建红.中国港口集装箱吞吐量预测:基于组合时间序列[J].系统科学与数学,2018,38(2):210-219.
[12]许利枝,汪寿阳.集装箱港口预测研究方法:香港港实证研究[J].管理科学学报,2015,18(5):46-56.
[13]梁小珍,乔晗,汪寿阳,等.基于奇异谱分析的我国航空客运量集成预测模型[J].系统工程理论与实践,2017,37(6):1479-1488.