基于多源异构时空数据融合的交通流量预测模型

doi:10.11959/j.issn.2096-0271.2023042

摘要/Abstract

摘要：

交通流量预测问题具有多源异构性，未来时刻的流量不仅与之前时刻的流量相关，同时也受城市区域间关系、天气情况、兴趣点（point of interest，POI）等异构时空数据的影响。针对此问题，提出一种基于多源异构时空数据融合的交通流量预测模型MHF-STNet。首先使用聚类方法获得城市区域不同的流量模式，并使用拼接、权重相加、注意力机制等多种方式融合交通流量、城市区域间的位置关系、天气、POI、工作日、假期多个模态的时空数据，使用深度学习方法对异构数据统一建模，预测未来时刻的交通流量。在北京出租车、纽约出租车和纽约自行车3个流量数据集上进行实验，与经典的交通流量预测模型相比，MHFSTNet的预测准确度有所提升。结果验证了MHF-STNet对异构时空数据统一建模的有效性。

关键词: 交通流量预测, 多源异构数据统一建模, 时空相关性

Abstract:

The problem of traffic flow forecasting has multi-source heterogeneity.The traffic flow in the future is not only related to the flow at the previous moment, but also affected by heterogeneous spatio-temporal data such as the relationship between urban regions, weather conditions and POI (point of interest).To solve this problem, a traffic flow prediction model based on multi-source heterogeneous spatio-temporal data fusion was proposed, which was called MHFSTNet (multi-source heterogeneous fusion spatio-temporal network).Firstly, this model used clustering methods to obtain different traffic patterns in urban areas, and utilized various methods such as concatenation, weight addition, and attention mechanism to integrate spatio-temporal data of multiple modalities, including traffic flow, location relationships between urban areas, weather, POI and the time of day.Deep learning methods were used to uniformly model heterogeneous data and predict traffic flow in the future.Experiments were conducted on three real-world traffic datasets, TaxiBJ, TaxiNYC and BikeNYC datasets.The results showed that MHF-STNet achieved the best performance compared with some classic traffic flow prediction models, which verified the effectiveness of MHF-STNet for unified modeling of heterogeneous spatio-temporal data.

Key words: traffic flow prediction, unified modeling of multi-source heterogeneous data, spatio-temporal correlation

中图分类号:

TP311

安洋, 孙健玮, 李倩, 宫永顺. 基于多源异构时空数据融合的交通流量预测模型[J]. 大数据, 2023, 9(4): 69-82.

Yang AN, Jianwei SUN, Qian LI, Yongshun GONG. Urban traffic flow prediction based on the multisource heterogeneous spatio-temporal data fusion[J]. Big Data Research, 2023, 9(4): 69-82.

图/表 11

图1

图2

图3

图4

图5

图6

图7

表1

表2

表3

表4

参考文献 23

[1]	STEPHANEDES Y J , MICHALOPOULOS P G , PLUM R A . Comparative performance evaluation of demand prediction algorithms[J]. Traffic Engineering and Control, 1981,22(10): 544-547.
[2]	WILLIAMS B M , HOEL L A . Modeling and forecasting vehicular traffic flow as a seasonal ARIMA process:theoretical basis and empirical results[J]. Journal of Transportation Engineering, 2003,129(6): 664-672.
[3]	VAN L , VAN H . Short-term traffic and travel time prediction models[J]. Artificial Intelligence Applications to Critical Transportation Issues, 2012,22(1): 22-41.
[4]	ZHANG J B , ZHENG Y , QI D K . Deep spatio-temporal residual networks for citywide crowd flows prediction[C]// Proceedings of the AAAI Conference on Artificial Intelligence. New York:ACM Press, 2017: 1655-1661.
[5]	KIPF T N , WELLING M . Semi-supervised classification with graph convolutional networks[C]// Proceedings of International Conference on Learning Representations.[S.l.:s.n.], 2017.
[6]	GUO S N , LIN Y F , FENG N ,et al. Attention based spatial-temporal graph convolutional networks for traffic flow forecasting[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019,33(1): 922-929.
[7]	VASWANI A , SHAZEER N , PARMAR N ,et al. Attention is all you need[J]. Advances in Neural Information Processing Systems, 2017,30.
[8]	GUO K , HU Y L , QIAN Z ,et al. Dynamic graph convolution network for traffic forecasting based on latent network of Laplace matrix estimation[J]. IEEE Transactions on Intelligent Transportation Systems, 2020,23(2): 1009-1018.
[9]	BAI L , YAO L N , LI C ,et al. Adaptive graph convolutional recurrent network for traffic forecasting[C]// Proceedings of the 34th International Conference on Neural Information Processing Systems. New York:ACM Press, 2020: 17804-17815.
[10]	何俊, 张彩庆, 李小珍 ,等. 面向深度学习的多模态融合技术研究综述[J]. 计算机工程, 2020,46(5): 1-11.
	HE J , ZHANG C Q , LI X Z ,et al. Survey of research on multimodal fusion technology for deep learning[J]. Computer Engineering, 2020,46(5): 1-11.
[11]	王冬海, 卢峰, 方晓蓉 ,等. 海洋大数据关键技术及在灾害天气下船舶行为预测上的应用[J]. 大数据, 2017,3(4): 81-90.
	WANG D H , LU F , FANG X R ,et al. Ocean big data and applications in ship behavior prediction under disaster weather[J]. Big Data Research, 2017,3(4): 81-90.
[12]	陈汐, 王印海, 代壮 ,等. 基于多源城市交通出行数据的定制公交需求辨识方法研究[J]. 大数据, 2020,6(6): 105-118.
	CHEN X , WANG Y H , DAI Z ,et al. Research on demand identification for customized bus based on multi-source mobility data[J]. Big Data Research, 2020,6(6): 105-118.
[13]	匡秋明, 杨雪冰, 张文生 ,等. 多源数据融合高时空分辨率晴雨分类[J]. 软件学报, 2017,28(11): 2925-2939.
	KUANG Q M , YANG X B , ZHENG W S ,et al. Fusion of multi-source data for rain/no-rain classification with high spatiotemporal resolution[J]. Journal of Software, 2017,28(11): 2925-2939.
[14]	JIN Z C , GONG T , YU D D ,et al. Mining contextual information beyond image for semantic segmentation[C]// Proceedings of the IEEE/CVF International Conference on Computer Vision. Piscataway:IEEE Press, 2021: 7231-7241.
[15]	BIANCHI F M , GRATTAROLA D , ALIPPI C . Spectral clustering with graph neural networks for graph pooling[C]// International Conference on Machine Learning. New York:ACM Press, 2020: 874-883.
[16]	张建晋, 王韫博, 龙明盛 ,等. 面向季节性时空数据的预测式循环网络及其在城市计算中的应用[J]. 计算机学报, 2020,43(2): 286-302.
	ZHANG J J , WANG Y B , LONG M S ,et al. Predictive recurrent networks for seasonal spatiotemporal data with applications to urban computing[J]. Chinese Journal of Computers, 2020,43(2): 286-302.
[17]	乔少杰, 韩楠, 岳昆 ,等. 基于数据场聚类的共享单车需求预测模型[J]. 软件学报, 2021,33(4): 1451-1476.
	QIAO S J , HAN N , YUE K ,et al. Sharedbike demand prediction model based on station clustering[J]. Journal of Software, 2022,33(4): 1451-1476.
[18]	冯宁, 郭晟楠, 宋超 ,等. 面向交通流量预测的多组件时空图卷积网络[J]. 软件学报, 2019,30(3): 759-769.
	FENG N , GUO S N , SONG C ,et al. Multi-component spatial-temporal graph convolution networks for traffic flow forecasting[J]. Journal of Software, 2019,30(3): 759-769.
[19]	LIANG Y X , OUYANG K , SUN J K ,et al. Fine-grained urban flow prediction[C]// Proceedings of the Web Conference. New York:ACM Press, 2021: 1833-1845.
[20]	LIN Z Q , FENG J , LU Z Y ,et al. DeepSTN+:context-aware spatialtemporal neural network for crowd flow prediction in metropolis[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019,33(1): 1020-1027.
[21]	HE K M , ZHANG X Y , REN S Q ,et al. Deep residual learning for image recognition[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2016: 770-778.
[22]	SCHUSTER M , PALIWAL K K . Bidirectional recurrent neural networks[J]. IEEE Transactions on Signal Processing, 1997,45(11): 2673-2681.
[23]	HOCHREUTER S , SCHMIDHUBER J . Long short-term memory[J]. Neural Computation, 1997,9(8): 1735-1780.

数据集	数据类型	时间范围	时间间隔	网格大小	外部特征
北京出租车数据集	出租车GPS数据	2013年7月1日—2013年10月30日	30 min	（32，32）	41种假期
（TaxiBJ）		2014年3月1日—2014年6月30日			16种天气
		2015年3月1日—2015年6月30日			14种POI
		2015年11月1日—2016年10月4日
纽约出租车数据	出租车轨迹数据	2015年1月1日—2015年3月1日	1h	（10，20）	20种假期
集（TaxiNYC）
纽约自行车数据	自行车租借数据	2014年4月1日—2014年9月30日	1h	（16，8）	20种假期
集（BikeNYC）

模型	北京出租车数据集		纽约出租车数据集		纽约自行车数据集
模型	MAE	RMSE	MAE	RMSE	MAE	RMSE
HA	16.70	30.76	9.22	29.19	3.77	8.61
ARIMA	17.36	23.66	8.38	17.97	3.07	5.24
RNN	11.16	17.97	5.85	12.66	2.57	5.04
LSTM	10.89	17.53	5.69	12.54	2.43	4.83
DeepSTN+	10.90	17.32	5.47	12.35	2.53	5.02
ST-ResNet	10.60	16.95	4.91	12.28	2.48	5.05
ASTGCN	10.70	17.25	5.67	12.89	2.44	4.84
MHF-STNet	10.55	16.68	4.82	11.98	2.36	4.68

时空数据	MAE	RMSE	对比模型	MAE	RMSE
流量数据	2.52	5.10	RNN	2.57	5.04
流量数据+POI	2.51	4.87	DeepSTN+	2.53	5.02
流量数据+天气、工作日、假期	2.45	4.95	ST-ResNet	2.48	5.05
流量数据+地理位置关系	2.40	4.78	ASTGCN	2.44	2.84
流量数据+交通模式	2.43	4.78	MHF-STNet没有软聚类	2.44	4.81

时空数据	MAE	RMSE
流量数据	2.52	5.10
流量数据+外部特征	2.45	4.86
流量数据+外部特征+交通模式	2.42	4.75
流量数据+外部特征+交通模式+地理位置关系	2.36	4.68