考虑货运网络港口间空间关联能提高短期预测精度，但是不同于公路网络中通过交叉路口形成简单明确的上下游空间关系，超高维的水运货运空间关系数据难以直接利用，为此，基于频繁港口和神经网络(FR-NN)构建考虑时空因素的水路货运量预测模型.该模型基于频繁模式的思想挖掘出目标港口在货运空间网络中的频繁港口，利用频繁港口提取货运网络的主要空间关系，用低维数据保留高维网络主要空间特征，再利用神经网络拟合频繁港口与目标港口货运量间的时空关系.实例分析表明，考虑时空关系有助于提高预测精度，模型能提高不同粒度的短期货运量预测精度，尤其是能够预测采用时间序列方法不理想的周、日货运量.
Considering spatial correlation in shipping network is an effective method to improve veracity of the short-term freight volume prediction. But the shipping network is much more complicated than the highway transportation. The high-dimensional data of shipping network cannot be used effectively. A short-term freight volume prediction model FR-NN which is based on frequent pattern mining and neural network is constructed. Frequent ports in cargo network are found out according to idea of frequent pattern mining, and they represent the main spatial relations in network. Neural network is used for fitting statistical relationships among ports. Example analysis shows that the model can improve veracity of the short-term freight volume prediction in different time particle sizes. Furthermore, the model can forecast to the weeks and days freight volume, that the time series method is not ideal for the prediction results.
［1］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.
［2］LI P. Research on forecasting model of cargo throughput of port group in the upper Yangtze River［C］//Proceedings of the 6th International Conference on Logistics and Supply Chain Management. Hong Kong: Aussino Academic Publishing House, 2016: 106-113.
［3］ZHA X, CHAI Y, WITLOX F, et al. Container throughput time series forecasting using a hybrid approach［J］. Lecture Notes in Electrical Engineering, 2016, 359: 639-650.
［4］GAO Y, LUO M, ZOU G. Forecasting with model selection or model averaging: A case study for monthly container port throughput［J］. Transportmetrica A: Transport Science, 2016, 12(4): 366-384.
［5］吴丽彬, 晏启鹏. 短期货运预测模型研究［J］. 交通运输工程与信息学报, 2008, 6(1): 64-69.
WU Libin, YAN Qipeng. Study of short-term freight prediction model ［J］. Journal of Transportation Engineering and Information, 2008, 6(1): 64-69.
［6］CHANG G, WANG S, XIAO X. Review of spatio-temporal models for short-term traffic forecasting［C］//2016 IEEE International Conference on Intelligent Transportation Engineering. Singapore: IEEE, 2016: 8-15.
［7］VLAHOGIANNI E I, KARLAFTIS M G, GOLIAS J. Spatio-temporal shortterm urban traffic volume forecasting using genetically optimized modular networks［J］. Computer-Aided Civil and Infrastructure Engineering, 2007, 22(5): 317-325.
［8］ZENG X, ZHANG Y. Development of recurrent neural network considering temporal-spatial input dynamics for freeway travel time modeling［J］. Computer-Aided Civil and Infrastructure Engineering, 2013, 28: 359-371.
［9］WANG J, TSAPAKIS I, ZHONG C. A space-time delay neural network model for travel time prediction［J］. Engineering Applications of Artificial Intelligence, 2016, 52: 145-160.
［10］田保慧, 郭彬. 基于时空特征分析的短时交通流预测模型［J］. 重庆交通大学学报(自然科学版), 2016, 35(3): 105-111.
TIAN Baohui, GUO Bin. A short-term traffic flow prediction model based on spatio-temporal characteristics analysis［J］. Journal of Chongqing Jiaotong University (Natural Science), 2016, 35(3): 105-111.
［11］HODGE V J, KRISHNAN R, AUSTIN J, et al. Short-term prediction of traffic flow using a binary neural network［J］. Neural Computing & Applications, 2014, 25(7/8): 1639-1655.
