| 基于数据分解的上海港集装箱吞吐量预测模型 |
| |
| 引用本文： | 冯宏祥,GRIFOLL Manel,AGUSTI Martinmallofre,郑彭军.基于数据分解的上海港集装箱吞吐量预测模型[J].中国航海,2019(2):132-138. |
| |
| 作者姓名： | 冯宏祥 GRIFOLL Manel AGUSTI Martinmallofre 郑彭军 |
| |
| 作者单位： | 宁波大学海运学院;港口贸易合作与发展协调创新中心港口经济协调创新中心;国家道路交通管理工程技术中心宁波大学分中心;加泰罗尼亚理工大学航海学院巴塞罗纳交通创新研究中心 |
| |
| 基金项目： | 浙江省科技厅公益技术应用研究项目(2016C31111);浙江省社会科学界联合会研究项目(2015N042);浙江省高等教育学会实验室工作研究重点项目(ZD201503) |
| |
| 摘 要： | 根据“分而治之”的框架,分别运用经验模式分解(Empirical Mode Decomposition, EMD)算法和季节性自回归积分滑动平均(Seasonal Autoregressive Integrated Moving Average, SARIMA)算法,将月度集装箱吞吐量时间系列数据分解为不同特征的分量,用支持向量回归(Support Vector Regression, SVR)模型分别对各分量进行预测,EMD-SVR模型和SARIMA-SVR模型预测结果的平均绝对百分误差(Mean Absolute Percentage Errors, MAPE)分别为 5.18%和7.26%,与港口实际吞吐量均较为一致,优于SVR模型的8.55%、自回归积分滑动平均(Autoregressive Integrated Moving Average, ARIMA)模型的11.8%和灰色系统(Grey Model, GM(1,1))模型的10.1%,验证数据分解方法在上海港集装箱月度吞吐量预测中的可行性,支持间接性预测模型精度高于直接模型的观点。
|
| 关 键 词： | 经验模式分解 支持向量回归 自回归积分滑动平均模型 灰色预测 |
Container Throughput Forecasting Model for Shanghai Port Based on Data Decomposition Method |
| |
| FENG Hongxiang,GRIFOLL Manel,AGUSTI Martinmallofre,ZHENG Pengjun.Container Throughput Forecasting Model for Shanghai Port Based on Data Decomposition Method[J].Navigation of China,2019(2):132-138. |
| |
| Authors: | FENG Hongxiang GRIFOLL Manel AGUSTI Martinmallofre ZHENG Pengjun |
| |
| Institution: | (Faculty of Maritime and Transportation,Ningbo University,Ningbo 315211,China;Collaborative Innovation of Port Economics,Centre for Collaborative Innovation on Port Trading Cooperation and Development,Ningbo 315211,China;Ningbo University Sub-Centre National Traffic Management Engineering & Technology Research Centre,Ningbo 315211,China;Barcelona Innovation in Transport (BIT),Barcelona School of Nautical Studies,Universitat Politècnica de Catalunya-BarcelonaTech,08003 Barcelona,Spain) |
| |
| Abstract: | The monthly container throughput time series are decomposed into components of different features by means of empirical mode decomposition EMD(Empirical Mode Decomposition) and seasonal SARIMA(Seasonal Autoregressive Integrated Moving Average), and the components are individually predicted with SVR(Support Vector Regression) model. The mean absolute percentage errors (MAPEs) with EMD-SVR and SARIMA-SVR are 5.18% and 7.26% respectively, which agrees with the actual data of Shanghai port better than that with SVR (8.55%), ARIMA(Autoregressive Integrated Moving Average)(11.8%) or GM(Grey Model)(1,1)(10.1%). The data decomposition method is proved, supporting that the higher precision shall be obtained in indirect prediction model. |
| |
| Keywords: | EMD ARIMA SVR grey forecasting |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
