基于NeuralProphet-LSTM组合模型的港口货物吞吐量预测

杨宇鸽; 郝杨杨; 王逸文

doi:10.3969/j.issn.1000-4653.2023.04.012

基于NeuralProphet-LSTM组合模型的港口货物吞吐量预测

上海海事大学物流研究中心, 上海 201306

基金项目:

上海市科技计划项目(23692106900)

详细信息

作者简介:
杨宇鸽(1999—),女,硕士研究生,研究方向为港口物流。E-mail:1124313129@qq.com

通讯作者:
郝杨杨(1984—),男,高级工程师,硕士生导师,研究方向为港航信息化与仿真优化。E-mail:yyhao@shmtu.edu.cn

中图分类号: U695.2
计量
- 文章访问数: 37
- HTML全文浏览量: 0
- PDF下载量: 6
出版历程
- 收稿日期: 2022-05-09

Prediction of port cargo throughput using NeuralProphet-LSTM combination model

Logistics Research Center, Shanghai Maritime University, Shanghai 201306, China

摘要

摘要: 为进一步提高货物吞吐量预测准确性，提出基于NeuralProphet时间序列模型与长短期记忆(LSTM)神经网络的组合预测模型。首先利用NeuralProphet模型对港口货物吞吐量数据进行训练得到预测值并计算残差序列，然后对残差数据建立LSTM神经网络模型进行预报修正，重构得到最终的预测值。以上海港、厦门港的月度货物吞吐量数据为样本展开试验，结果表明，该模型能够有效地解决数据异常波动造成的预测结果误差大、预测效果不稳定等问题；相比于传统单一模型与LSTM-支持向量机(SVM)、Bi-LSTM等组合模型，NeuralProphet-LSTM模型预测精度更高，可帮助港航企业及时调整规划决策与经营策略。
- 港口吞吐量 /
- 组合模型 /
- 预测 /
- NeuralProphet
Abstract: Conventional method has not been successful in predicting cargo throughput of ports due to the great fluctuation caused by complex external factors. In order to improve the accuracy of cargo throughput prediction for ports, the combination of NeuralProphet time series model and LSMT neural network is introduced. NeuralProphet model is trained to produce, according to input cargo throughput data, a preliminary prediction with a residual sequence. The LSMT model is used to correct the preliminary prediction and output final prediction. The combination model is verified through process the monthly throughput data from Shanghai Port and Xiamen Port. The advantage of the combination model in handling abnormal fluctuation over other configurations is demonstrated.
- port throughput /
- combination model /
- prediction /
- NeuralPropht

HTML全文

参考文献(17)

[1]	王茁,高璐.港口集装箱吞吐量预测方法研究[J].上海工程技术大学学报,2020,34(2):201-206.WANG Z,GAO L.Research on forecasting method of container throughput[J].Journal of Shanghai University of Engineering Science,2020,34(2):201-206.(in Chinese)
[2]	杨金花,杨艺.基于灰色模型的上海港集装箱吞吐量预测[J].上海海事大学学报,2014,35(2):28-32.YANG J H,YANG Y.Forecast of Shanghai Port container throughput based on grey model[J].Journal of Shanghai Maritime University,2014,35(2):28-32.(in Chinese)
[3]	吴琛.基于时间序列模型的港口集装箱吞吐量预测[J].珠江水运,2019(5):73-74.WU C.Based on the port container throughput forecast of time series model[J].Transportation Research,2019(5):73-74.(in Chinese)
[4]	刘枚莲,朱美华.基于BP神经网络的港口吞吐量预测模型[J].系统科学学报,2012,20(4):88-91.LIU M L,ZHU M H.Port throughput prediction model based on BP neural network[J].Journal of systems science,2012,20(4):88-91.(in Chinese)
[5]	李广儒,朱庆辉.基于Elman神经网络的港口货物吞吐量预测[J].重庆交通大学学报(自然科学版),2020,39(6):8-12.LI G R,ZHU Q H.Prediction of port cargo throughput based on Elman neural network[J].Journal of Chongqing Jiaotong University (Natural Science Edition),2020,39(6):8-12.(in Chinese)
[6]	王凤武,张晓博,阎际驰,等.基于LSTM的上海港集装箱吞吐量预测[J].中国航海,2022,45(2):109-114. WANG F W,ZHANG X B,YAN J C,et al.Container throughput forecast of Shanghai Port based on LSTM[J].Navigation of China,2022,45(2):109-114.
[7]	王振振,苌道方,朱宗良,罗天.基于ES-Markov模型的港口集装箱季度吞吐量分析与预测[J].中国航海,2019,42(4):125-130.WANG Z Z,CHANG D F,ZHU Z L,et al.Analysis and prediction of quarterly container throughput based on ES-Markov model[J].China navigation,2019,42(4):125-130.(in Chinese)
[8]	张树奎,肖英杰,鲁子爱.基于灰色神经网络的港口集装箱吞吐量预测模型研究[J].重庆交通大学学报(自然科学版),2015,34(5):135-138.ZHANG S K,XIAO Y J,LU Z A.Research on prediction model of port container throughput based on grey neural network[J].Journal of Chongqing Jiaotong University (Natural Science Edition),2015,34(5):135-138.(in Chinese)
[9]	张丰婷,杨菊花,任金荟,金坤.基于优化变分模态分解和核极限学习机的集装箱吞吐量预测[J/OL].计算机应用:1-10[2022-07-09].ZHANG F T,YANG J H,REN J H,et al.Container throughput prediction based on optimized variational mode decomposition and kernel extreme learning machine[J].Computer Applications:1-10[2022-07-09].(in Chinese)
[10]	顾子瑜,陈诺.基于ICAKELM的港口集装箱吞吐量预测模型[J].中国航海,2022,45(2):93-99. GU Z Y,CHEN N.Port container throughput forecasting model based on ICAKELM[J].Navigation of China,2022,45(2):93-99.
[11]	XIE G,WANG S,ZHAO Y,et al.Hybrid approaches based on LSSVR model for container throughput forecasting:a comparative study[J].Applied Soft Computing,2013,13(5):2232-2241.
[12]	TRIEBE O,HEWAMALAGE H,PILYUGINA P,et al.Neuralprophet:explainable forecasting at scale[J].arXiv preprint arXiv:2111.15397,2021.
[13]	吴晓峰,杨颖梅,陈垚彤.基于BP神经网络误差校正的ARIMA组合预测模型[J].统计与决策,2019,35(15):65-68.WU X F,YANG Y M,CHEN Y T.ARIMA combination forecast based on BP neural network error correction model[J].Journal of Statistics and Decision,2019,35(15):65-68.(in Chinese)
[14]	凌立文,张大斌.组合预测模型构建方法及其应用研究综述[J].统计与决策,2019,35(1):18-23.LING L W,ZHANG D B.Combination forecast model building method and its application research review[J].Journal of Statistics and Decision,2019,35(1):18-23.(in Chinese)
[15]	靳廉洁,沈益华,高天航,等.中美贸易摩擦对我国沿海港口集装箱吞吐量影响研究[J].中国航海,2022,45 (1):73-78. JIN L J,SHEN Y H,GAO T H,et al.Study on the impact of Sino-US trade friction on container throughput of China's coastal ports[J].Navigation of China,2022,45(1):73-78.
[16]	熊红林,冀和,樊重俊,杨梦达.基于LSTM-SVR模型的航空旅客出行指数预测[J].系统管理学报,2020,29(6):1169-1176.XIONG H L,JI H,FAN C J,et al.Forecasting air passenger travel index based on LSTM-SVR model[J].Journal of systems management,2020,29(6):1169-1176.(in Chinese)
[17]	孙晓聪,付玉慧.基于RF-双向LSTM的集装箱吞吐量预测[J].上海海事大学学报,2022,43(1):60-65.SUN X C,FU Y H.Based on Bi-directional RF - LSTM container throughput forecast[J].Journal of Shanghai Maritime University,2022,43(1):60-65.(in Chinese)