浏览全部资源
扫码关注微信
纸质出版日期： 2021-05-16 ，
录用日期： 2020-11-11
移动端阅览
杨成意, 刘峰, 齐佳音, 段妍, 吕润倩, 肖子龙. 面向COVID-19疫情预测的图卷积神经网络时空数据学习[J]. 中国图象图形学报, 2021,26(5):1128-1137.
Chengyi Yang, Feng Liu, Jiayin Qi, Yan Duan, Runqian Lyu, Zilong Xiao. Spatiotemporal data learning of graph convolutional neural network for epidemic prediction of COVID-19[J]. Journal of Image and Graphics, 2021,26(5):1128-1137.
目的
2
当前的疾病传播研究主要集中于时序数据和传染病模型，缺乏运用空间信息提升预测精度的探索和解释。在处理时空数据时需要分别提取时间特征和空间特征，再进行特征融合得到较为可靠的预测结果。本文提出一种基于图卷积神经网络（graph convolutional neural network，GCN）的时空数据学习方法，能够运用空间模型端对端地学习时空数据，代替此前由多模块单元相集成的模式。
方法
2
依据数据可视化阶段呈现出的地理空间、高铁线路、飞机航线与感染人数之间的正相关关系，将中国各城市之间的空间分布关系和交通连接关系映射成网络图并编码成地理邻接矩阵、高铁线路直达矩阵、飞机航线直达矩阵以及飞机航线或高铁线路直达矩阵。按滑动时间窗口对疫情数据进行切片后形成张量，依次分批输入到图深度学习模型中参与卷积运算，通过信息传递、反向传播和梯度下降更新可训练参数。
结果
2
在新型冠状病毒肺炎疫情数据集上的实验结果显示，采用GCN学习这一时空数据的分布特征相较于循环神经网络模型，在训练过程中表现出了更强的拟合能力，在训练时间层面节约75%以上的运算成本，在两类损失函数下的平均测试集损失能够下降80%左右。
结论
2
本文所采用的时空数据学习方法具有较低的运算成本和较高的预测精度，尤其在空间特征强于时间特征的时空数据中有着更好的性能，并且为流行病传播范围和感染人数的预测提供了新的方法和思路，有助于相关部门在公共卫生事件中制定应对措施和疾病防控决策。
Objective
2
COVID-19 has caused a severe impact on the medical system and economic growth of countries all over the world. Therefore
the epidemic information of each city has important reference value for governments and enterprises to formulate public health prevention and control measures and decisions in opening the economy. According to relevant research
the infectious disease model and time series model have played an important role in finding potential hosts
confirming human to human transmission
and estimating the basic reproductive number. Related research methods on disease transmission and confirmed case prediction have experienced a series of evolution
including demographic method
dynamic model
social network analysis
flight passenger volume estimation
data mining
and machine learning method. However
the prediction accuracy of these methods still needs to be improved by using spatial information in the study of epidemic transmission. In recent years
the boom of graph deep learning has provided new technologies and methods for the estimation of epidemic spread. From the iteration of trainable parameters in the way of information interaction in early time to the optimization of graph type
propagation mechanism
and output steps
this development process laid the foundation for the generation of graph convolutional neural network(GCN). The development of graph convolution network optimizes the performance of graph neural network in spectral domain and spatial domain by changing convolution kernel and information aggregation mode. The progress of representation learning improves the convenience of graph data processing. The rise of integrated framework realizes more accurate prediction in spatiotemporal data processing represented by traffic flow.
Method
2
Compared with traffic flow prediction
epidemic data prediction has stronger spatial attribute and weaker temporal attribute
which is the reason why GCN is used alone instead of integration approaches. First
according to the data visualization stage of the epidemic information and geographical location and traffic network between the positive correlation
the spatial distribution relationship and traffic connection relationship between affected cities in China are mapped into a graph network and encoded into geographic adjacency matrix
high-speed railway direct matrix
aircraft route direct matrix
and aircraft route or high-speed railway direct matrix. Four cities networks with different connection modes are formed by these four adjacent matrices
and the corresponding GCN model is constructed based on these cities networks
including geographical proximity graph convolutional neural network(GPGCN)
airline graph convolutional neural network(ALGCN)
high speed railway graph convolutional neural network(HSRGCN)
airline and high speed railway graph convolutional neural network(ALHSRGCN). After dividing the training
validation
and test sets at a ratio of 6:2:2
the epidemic data were sliced according to the sliding time window to form a three-dimensional tensor with a size of 30×327×7 as the test set
which was input into the graph deep learning model in batches to participate in convolution operation. The training parameters were updated by information transmission
back propagation
and gradient descent.
Result
2
The experimental results on the COVID-19 dataset demonstrated that the learning distribution features of spatiotemporal data by the GCN model showed stronger fitting ability than the recurrent neural network model in the training process. It could save more than 75% of the computation cost at the training time level
and the average test set loss on mean absolute error(MAE) and mean square error(MSE) decreased by about 80%. The value of loss would converge to a lower position and achieve a more stable training process so as to obtain more accurate prediction when MSE is chosen as the loss function.
Conclusion
2
The spatiotemporal data learning method in this study has lower operation cost and higher prediction accuracy
which shows better performance especially in the case of spatiotemporal data with stronger spatial characteristics than temporal characteristics. It provides new approaches and ideas for the prediction of epidemic spread range and number of infected people
which is conducive for relevant departments to formulate countermeasures for disease prevention and control decisions in public health events.
深度学习图卷积神经网络(GCN)时空数据处理新冠肺炎疫情预测
deep learninggraph convolutional neural network(GCN)spatiotemporal data processingCoronavirus Disease 2019(COVID-19)epidemic prediction
Beck D, Haffari G and Cohn T. 2018. Graph-to-sequence learning using gated graph neural networks//Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics. Melbourne, Australia: ACL: 273-283[DOI: 10.18653/v1/P18-1026http://dx.doi.org/10.18653/v1/P18-1026]
Brockmann D and Helbing D. 2013. The hidden geometry of complex, network-driven contagion phenomena. Science, 2013, 342(6164): 1337-1342[DOI: 10.1126/science.1245200http://dx.doi.org/10.1126/science.1245200]
Bruna J, Zaremba W, Szlam A and LeCun Y. 2014. Spectral networks and locally connected networks on graphs[EB/OL]. [2020-09-03].https://arxiv.org/pdf/1312.6203.pdfhttps://arxiv.org/pdf/1312.6203.pdf
Chao D L, Halloran M E, Obenchain V J and Longini Jr I M. 2010. FluTE, a publicly available stochastic influenza epidemic simulation model. PLoS Computational Biology, 2010, 6(1): #e1000656[DOI: 10.1371/journal.pcbi.1000656http://dx.doi.org/10.1371/journal.pcbi.1000656]
Chen Y, Cheng J, Jiang Y and Liu K J. 2020. A time delay dynamical model for outbreak of 2019-nCoV and the parameter identification[EB/OL]. [2020-09-13].https://arxiv.org/pdf/2002.00418.pdfhttps://arxiv.org/pdf/2002.00418.pdf
Du J, Zhang S H, Wu G H, Moura J M F and Kar S. 2017. Topology adaptive graph convolutional networks[EB/OL]. [2020-09-03].https://arxiv.org/pdf/1710.10370.pdfhttps://arxiv.org/pdf/1710.10370.pdf
Duvenaud D, Maclaurin D, Aguilera-Iparraguirre J, Gómez-Bombarelli R, Hirzel T, Aspuru-Guzik A and Adams R. 2015. Convolutional networks on graphs for learning molecular fingerprints//Proceedings of the 28th International Conference on Neural Information Processing Systems. Montreal, Canada: NIPS: 2224-2232[DOI: 10.5555/2969442.2969488http://dx.doi.org/10.5555/2969442.2969488]
Fang X Y, Liu W D, Ai J, He M K, Wu Y, Shi Y Y, Shen W Q and Bao C J. 2020. Forecasting incidence of infectious diarrhea using random forest in Jiangsu Province, China. BMC Infectious Diseases, 2020, 20(1): #222[DOI: 10.1186/s12879-020-4930-2http://dx.doi.org/10.1186/s12879-020-4930-2]
Fusco T, Bi Y X, Wang H Y and Browne F. 2020. Data mining and machine learning approaches for prediction modelling of schistosomiasis disease vectors: epidemic disease prediction modelling. International Journal of Machine Learning and Cybernetics, 11(6): 1159-1178[DOI:10.1007/s13042-019-01029-x]
Geng X, Li Y G, Wang L Y, Zhang L Y, Yang Q, Ye J P and Liu Y. 2019. Spatiotemporal multi-graph convolution network for ride-hailing demand forecasting//Proceedings of the 33rd AAAI Conference on Artificial Intelligence. Honolulu, USA: AAAI: 3656-3663[DOI: 10.1609/aaai.v33i01.33013656http://dx.doi.org/10.1609/aaai.v33i01.33013656]
Glorot X and Bengio Y. 2010. Understanding the difficulty of training deep feedforward neural networks//Proceedings of the 13th International Conference on Artificial Intelligence and Statistics. Sardinia, Italy: JMLR. org: 249-256
Gori M, Monfardini G and Scarselli F. 2005. A new model for learning in graph domains//Proceedings of 2005 IEEE International Joint Conference on Neural Networks. Montreal, Canada: IEEE: 729-734[DOI: 10.1109/IJCNN.2005.1555942http://dx.doi.org/10.1109/IJCNN.2005.1555942]
Guo Q, Li M, Wang C H, Wang P H, Fang Z C, Tan J, Wu S F, Xiao Y H and Zhu H Q. 2020. Host and infectivity prediction of Wuhan 2019 novel coronavirus using deep learning algorithm[EB/OL]. [2020-01-21].https://www.biorxiv.org/content/10.1101/2020.01.21.914044v2.full.pdfhttps://www.biorxiv.org/content/10.1101/2020.01.21.914044v2.full.pdf
Guo S N, Lin Y F, Feng N, Song C and Wan H Y. 2019. Attention based spatial-temporal graph convolutional networks for traffic flow forecasting//Proceedings of the 33rd AAAI Conference on Artificial Intelligence. Honolulu, USA: AAAI: 922-929[DOI: 10.1609/aaai.v33i01.3301922http://dx.doi.org/10.1609/aaai.v33i01.3301922]
Hamilton W L, Ying R and Leskovec J. 2017. Inductive representation learning on large graphs//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, USA: NIPS: 1025-1035[DOI: 10.5555/3294771.3294869http://dx.doi.org/10.5555/3294771.3294869]
Kampffmeyer M, Chen Y B, Liang X D, Wang H, Zhang Y J and Xing E P. 2019. Rethinking knowledge graph propagation for zero-shot learning//Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach, USA: IEEE: 11479-11488[DOI: 10.1109/CVPR.2019.01175http://dx.doi.org/10.1109/CVPR.2019.01175]
Kipf T N and Welling M. 2016. Semi-supervised classification with graph convolutional networks[EB/OL]. [2020-08-18].https://arxiv.org/pdf/1609.02907v4.pdfhttps://arxiv.org/pdf/1609.02907v4.pdf
Li Q, Guan X H, Wu P, Wang X Y, Zhou L, Tong Y Q, Ren R Q, Leung K S M, Lau E H Y, Wong J Y, Xing X S, Xiang N J, Wu Y, Li C, Chen Q, Li D, Liu T, Zhao J, Liu M, Tu W X, Chen C D, Jin L M, Yang R, Wang Q, Zhou S H, Wang R, Liu H, Luo Y B, Liu Y, Shao G, Li H, Tao Z F, Yang Y, Deng Z Q, Liu B X, Ma Z T, Zhang Y P, Shi G Q, Lam T T Y, Wu J T, Gao G F, Cowling B J, Yang B, Leung C M and Feng Z J. 2020. Early transmission dynamics in Wuhan, China, of novel coronavirus-infected pneumonia. New England Journal of Medicine, 382(13): 1199-1207[DOI:10.1056/NEJMoa2001316]
Li R Y, Wang S, Zhu F Y and Huang J Z. 2018. Adaptive graph convolutional neural networks//Proceedings of the 32nd AAAI Conference on Artificial Intelligence. New Orleans, USA: AAAI: 3546-3553
Li Y J, Tarlow D, Brockschmidt M and Zemel R. 2015. Gated graph sequence neural networks[EB/OL]. [2020-08-18].https://arxiv.org/pdf/1511.05493.pdfhttps://arxiv.org/pdf/1511.05493.pdf
Manessi F, Rozza A and Manzo M. 2020. Dynamic graph convolutional networks. Pattern Recognition, 97: #107000[DOI:10.1016/j.patcog.2019.107000]
May R M, Anderson R M and McLean A R. 1988. Possible demographic consequences of HIV/AIDS epidemics. I. Assuming HIV infection always leads to AIDS. Mathematical Biosciences, 90(1-2): 475-505[DOI:10.1016/0025-5564(88)90079-X]
Merler S, Ajelli M, Pugliese A and Ferguson N M. 2011. Determinants of the spatiotemporal dynamics of the 2009 H1 N1 pandemic in Europe: implications for real-time modelling. PLoS Computational Biology, 7(9): #e1002205[DOI:10.1371/journal.pcbi.1002205]
Monti F, Boscaini D, Masci J, RodolàE, Svoboda J and Bronstein M M. 2017. Geometric deep learning on graphs and manifolds using mixture model CNNs//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, USA: IEEE: 5425-5434[DOI: 10.1109/CVPR.2017.576http://dx.doi.org/10.1109/CVPR.2017.576]
Niepert M, Ahmed M, Kutzkov K. 2016. Learning convolutional neural networks for graphs//Proceedings of the 33rd International Conference on International Conference on Machine Learning (ICML). New York, USA: JMLR: 2014-2023. [DOl: abs/10.5555/3045390.3045603http://dx.doi.org/abs/10.5555/3045390.3045603]
Scarselli F, Gori M, Tsoi A C, Hagenbuchner M and Monfardini G. 2009. The graph neural network model. IEEE Transactions on Neural Networks, 20(1): 61-80[DOI:10.1109/TNN.2008.2005605]
Trivedi R, Farajtabar M, Biswal P and Zha H Y. 2019. Dyrep: learning representations over dynamic graphs//Proceedings of the 7th International Conference on Learning Representations. New Orleans, USA: OpenReview. net: 1-25
Wu J T, LeungK and Leung G M. 2020. Nowcasting and forecasting the potential domestic and international spread of the 2019-nCoV outbreak originating in Wuhan, China: a modelling study. The Lancet, 395(10225): 689-697[DOI:10.1016/S0140-6736(20)30260-9]
Wu Z H, Pan S R, Chen F W, Long G D, Zhang C Q and Yu P S. 2021. A comprehensive survey on graph neural networks. IEEE Transactions on Neural Networks and Learning Systems, 32(1): 4-24[DOI:10.1109/TNNLS.2020.2978386]
Xu X T, Chen P, Wang J F, Feng J N, Zhou H, Li X, Zhong W and Hao P. 2020. Evolution of the novel coronavirus from the ongoing Wuhan outbreak and modeling of its spike protein for risk of human transmission. Science China Life Sciences, 63(3): 457-460[DOI:10.1007/s11427-020-1637-5]
Yu B, Yin H T and Zhu Z X. 2018. Spatio-temporal graph convolutional networks: a deep learning framework for traffic forecasting//Proceedings of the 27th International Joint Conference on Artificial Intelligence. Stockholm, Sweden: IJCAI: 3634-3640[DOI: 10.24963/ijcai.2018/505http://dx.doi.org/10.24963/ijcai.2018/505]
Zhang Y Z, Xiong Y, Kong X N, Li S S, Li J H and Zhu Y Y. 2018. Deep collective classification in heterogeneous information networks//Proceedings of 2018 World Wide Web Conference. Lyon, Framce: ACM: 399-408[DOI: 10.1145/3178876.3186106http://dx.doi.org/10.1145/3178876.3186106]
Zhao L, Song Y J, Zhang C, Liu Y, Wang P, Lin T, Deng M and Li H F. 2020. T-GCN: a temporal graph convolutional network for traffic prediction. IEEE Transactions on Intelligent Transportation Systems, 21(8): 3848-3858[DOI:10.1109/TITS.2019.2935152]
Zhou J, Cui G Q, Zhang Z Y, Yang C, Liu Z Y, Wang L F, Li C C and Sun M S. 2018. Graph neural networks: a review of methods and applications[EB/OL]. [2020-08-18].https://arxiv.org/pdf/1812.08434.pdfhttps://arxiv.org/pdf/1812.08434.pdf
Zhou T, Liu Q H, Yang Z M, Liao J Y, Yang K X, Bai W, Lu X and Zhang W. 2020. Preliminary prediction of the basic reproduction number of the Wuhan novel coronavirus 2019-nCoV. Journal of Evidence-Based Medicine, 13(1): 3-7[DOI:10.1111/jebm.12376]
相关作者
相关机构
京公网安备11010802024621