Prediction model for regional freight volume on highways based on spatiotemporal information fusion

doi:10.3785/j.issn.1008-973X.2025.10.010

Journal of ZheJiang University (Engineering Science)

2025, Vol. 59

Issue (10): 2096-2105 DOI: 10.3785/j.issn.1008-973X.2025.10.010

Prediction model for regional freight volume on highways based on spatiotemporal information fusion

Liying ZHAO1(

),Zhanzhong WANG2,*(

)

1. School of Economics and Management, Xi’an University of Technology, Xi’an 710054, China
2. Transportation College, Jilin University, Changchun 130022, China

Download:

HTML

PDF(1204KB) HTML
Export: BibTeX | EndNote (RIS)

Abstract

The freight volume data between regions of highway has mutual influence, and the temporal and spatial problems cannot be handled simultaneously by traditional long short-term memory (LSTM) models. An improved LSTM model (TS-LSTM) based on spatiotemporal information fusion was designed, and a method for reconstructing the dataset was proposed according to the importance of spatiotemporal information. To verify the effectiveness of the model, the highway toll system data (25 563 256 in total) for a consecutive 12-month period in a certain region was used as the original dataset, and TS-LSTM was compared and analyzed with a time-based LSTM model (T-LSTM), a space-based LSTM model (S-LSTM), a fully connected neural network, an unidirectional LSTM, a bidirectional LSTM, and the Transformer. Results showed that the performance of TS-LSTM varied across different regions, and compared to other machine learning models, the reduction range of the mean absolute error was between 40% and 85%. The mean absolute error of TS-LSTM was 10% lower than that of Transformer, and the mean absolute percentage error was 21 percentage points lower. The prediction performance of TS-LSTM were superior to those of the comparison model.

Key words： highway transportation regional freight volume forecast spatiotemporal information fusion long short-term memory (LSTM) model dataset reconstruction

Received: 23 August 2024 Published: 27 October 2025

CLC:

U 492.3

Fund: 陕西省社会科学基金资助项目（2021R025）；陕西省自然科学基础研究计划（2024JC-YBMS-376）；陕西省教育厅科学研究计划项目（23JK0557）.

Corresponding Authors: Zhanzhong WANG E-mail: lyzhao@xaut.edu.cn;wangzz@jlu.edu.cn

	Service
	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Liying ZHAO
	Zhanzhong WANG

Cite this article:

Liying ZHAO,Zhanzhong WANG. Prediction model for regional freight volume on highways based on spatiotemporal information fusion. Journal of ZheJiang University (Engineering Science), 2025, 59(10): 2096-2105.

URL:

https://www.zjujournals.com/eng/10.3785/j.issn.1008-973X.2025.10.010 OR https://www.zjujournals.com/eng/Y2025/V59/I10/2096

基于时空信息融合的高速公路区域货运量预测模型

高速公路区域间货运量数据相互影响，传统长短期记忆网络（LSTM）模型无法同时处理时空问题，为此设计基于时空信息融合的改进LSTM模型(TS-LSTM)，提出按照时空信息的重要性对数据集进行重构的方法. 为了验证模型有效性，以某地区连续12个月的高速公路收费系统数据(共25563256条)为原始数据集，将TS-LSTM与基于时间的LSTM模型（T-LSTM）、基于空间的LSTM模型（S-LSTM），全连接神经网络、单向LSTM、双向LSTM和Transformer进行对比分析. 结果表明，不同区域使用TS-LSTM效果不同，相比其他机器学习模型，TS-LSTM的平均绝对值误差降低范围在40%~85%；TS-LSTM的平均绝对值误差相比Transformer低10%，平均绝对百分比误差低21个百分点；TS-LSTM的预测效果均优于对比模型.

关键词： 高速公路运输, 区域货运量预测, 时空信息融合, 长短期记忆网络（LSTM）模型, 数据集重构

Fig.1 Structure diagram of classic long short-term memory

Fig.2 Structure diagram of improved long short-term memory model based on spatiotemporal information fusion

Fig.3 Cell internal structure diagram of improved long short-term memory model based on spatiotemporal information fusion

Fig.4 Results of temporal importance analysis for representative grids

Fig.5 Results of spatial importance analysis for representative grids

Tab.1 Comparison of freight volume prediction results among different long short-term memory models

Fig.6 Comparison chart of annual cyclical freight volume changes predicted by different long short-term memory models

Fig.7 Comparison of prediction performance among different long short-term memory models

Tab.2 Comparison of performance metrics for different prediction methods

Fig.8 Comparison of prediction performance of different prediction models on representative grids

Fig.9 Comparison of generalization capabilities between two prediction models


[1]	徐菲, 任爽基于分解—集成的铁路货运需求预测研究[J]. 运筹与管理, 2021, 30 (8): 133- 138 XU Fei, REN Shuang Railway freight demand forecasting based on decompose-ensemble method[J]. Operations Research and Management Science, 2021, 30 (8): 133- 138

[2]	蔡志强. 阿拉山口铁路货运量预测分析 [D]. 兰州: 兰州交通大学, 2015: 1–44. CAI Zhiqiang. The forecasting analysis of volume of railway freight about Ala Shan Kou [D]. Lanzhou: Lanzhou Jiaotong University, 2015: 1–44.

[3]	付建飞, 安仲文, 宋小满, 等基于灰色模型的铁路分品类货运量预测[J]. 交通运输工程与信息学报, 2014, 12 (3): 38- 42 FU Jianfei, AN Zhongwen, SONG Xiaoman, et al Traffic volume forecast of railway different freight categories based on grey model[J]. Journal of Transportation Engineering and Information, 2014, 12 (3): 38- 42

[4]	王开鹏. 基于高维面板数据的铁路货运量预测方法 [D]. 武汉: 华中科技大学, 2019: 1–87. WANG Kaipeng. Railway freight volume forecasting method based on high-dimensional panel data [D]. Wuhan: Huazhong University of Science and Technology, 2019: 1–87.

[5]	徐曼, 陆芬基于组合预测模型的陕西省物流需求预测[J]. 物流科技, 2023, 46 (11): 27- 31 XU Man, LU Fen Forecasting and analysis of Shaanxi Province’s logistics demand based on combination forecasting model[J]. Logistics Sci-Tech, 2023, 46 (11): 27- 31

[6]	席少龙. 基于PCA-AFSA-SVM的河南省公路货运量预测研究 [D]. 深圳: 深圳大学, 2019: 1–67. XI Shaolong. Prediction of highway freight volume in Henan Province based on PCA-AFSA-SVM [D]. Shenzhen: Shenzhen University, 2019: 1–67.

[7]	SOBASZEK Ł, GOLA A, KOZŁOWSKI E Predictive scheduling with Markov chains and ARIMA models[J]. Applied Sciences, 2020, 10 (17): 6121 doi: 10.3390/app10176121

[8]	NUZZOLO A, COMI A Urban freight demand forecasting: a mixed quantity/delivery/vehicle-based model[J]. Transportation Research Part E: Logistics and Transportation Review, 2014, 65: 84- 98 doi: 10.1016/j.tre.2013.12.014

[9]	GAO N, HONG Y, CHEN J, et al Regional expressway freight volume prediction algorithm based on meteorological information[J]. Promet - Traffic and Transportation, 2023, 35 (2): 195- 211

[10]	LI Y, WEI Z Regional logistics demand prediction: a long short-term memory network method[J]. Sustainability, 2022, 14 (20): 13478 doi: 10.3390/su142013478

[11]	MOHAMMADZADEH M, CHOUPANI A A, AFSHAR F The short-term prediction of daily traffic volume for rural roads using shallow and deep learning networks: ANN and LSTM[J]. The Journal of Supercomputing, 2023, 79 (15): 17475- 17494 doi: 10.1007/s11227-023-05333-w

[12]	YIN S, JIANG Y, TIAN Y, et al A data-driven fuzzy information granulation approach for freight volume forecasting[J]. IEEE Transactions on Industrial Electronics, 2017, 64 (2): 1447- 1456 doi: 10.1109/TIE.2016.2613974

[13]	YANG F, TANG X, GAN Y, et al Forecast of freight volume in Xi’an based on gray GM (1, 1) model and Markov forecasting model[J]. Journal of Mathematics, 2021, 2021: 6686786

[14]	CAI P, WANG Y, LU G, et al A spatiotemporal correlative k-nearest neighbor model for short-term traffic multistep forecasting[J]. Transportation Research Part C: Emerging Technologies, 2016, 62: 21- 34 doi: 10.1016/j.trc.2015.11.002

[15]	WU S, YANG Z, ZHU X, et al Improved k-nn for short-term traffic forecasting using temporal and spatial information[J]. Journal of Transportation Engineering, 2014, 140 (7): 04014026 doi: 10.1061/(ASCE)TE.1943-5436.0000672

[16]	TIAN Y, ZHANG K, LI J, et al LSTM-based traffic flow prediction with missing data[J]. Neurocomputing, 2018, 318: 297- 305 doi: 10.1016/j.neucom.2018.08.067

[17]	ZOU Y, ZHU X, ZHANG Y, et al A space–time diurnal method for short-term freeway travel time prediction[J]. Transportation Research Part C: Emerging Technologies, 2014, 43: 33- 49 doi: 10.1016/j.trc.2013.10.007

[1]	Gen LI,Wei ZHAI,Lan WU. Study of merging interactions based on gradient boosting decision tree[J]. Journal of ZheJiang University (Engineering Science), 2022, 56(4): 649-655.

Viewed

Full text

Abstract

Cited

Shared

Discussed