Abstract: Human mobility trajectories are fundamental resources for analyzing mobile behaviors in urban computing applications. However, these trajectories, typically collected from location-based services, often suffer from sparsity and irregularity in time. To support the development of mobile applications, there is a need to recover or estimate missing locations of unobserved time slots in these trajectories at a fine-grained spatial–temporal resolution. Existing methods for trajectory recovery rely on either individual user trajectories or collective mobility patterns from all users. The potential to combine individual and collective patterns for precise trajectory recovery remains unexplored. Additionally, current methods are sensitive to the heterogeneous temporal distributions of the observable trajectory segments. In this paper, we propose CLMove (where CL stands for contrastive learning), a novel model designed to capture multilevel mobility patterns and enhance robustness in trajectory recovery. CLMove features a two-stage location encoder that captures collective and individual mobility patterns. The graph neural network based networks in CLMove explore location transition patterns within a single trajectory and across various user trajectories. We also design a trajectory-level contrastive learning task to improve the robustness of the model. Extensive experimental results on three representative real-world datasets demonstrate that our CLMove model consistently outperforms state-of-the-art methods in terms of trajectory recovery accuracy.

基于对比学习的移动轨迹准确恢复

[1]Chandio AA, Tziritas N, Zhang F, et al., 2016. Towards adaptable and tunable cloud-based map-matching strategy for GPS trajectories. Front Inform Technol Electron Eng, 17(12):1305-1319.

[2]Chen GS, Viana AC, Fiore M, et al., 2019. Complete trajectory reconstruction from sparse mobile phone data. EPJ Data Sci, 8(1):30.

[3]Chen T, Kornblith S, Norouzi M, et al., 2020. A simple framework for contrastive learning of visual representations. Proc 37^th Int Conf on Machine Learning, p.1597-1607.

[4]Chen XL, He KM, 2021. Exploring simple Siamese representation learning. Proc IEEE/CVF Conf on Computer Vision and Pattern Recognition, p.15745-15753.

[5]Cho E, Myers SA, Leskovec J, 2011. Friendship and mobility: user movement in location-based social networks. Proc 17^th ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining, p.1082-1090.

[6]Chondrogiannis T, Bornholdt J, Bouros P, et al., 2022. History oblivious route recovery on road networks. Proc 30^th Int Conf on Advances in Geographic Information Systems, Article 44.

[7]Chung J, Gülçehre Ç, Cho K, et al., 2014. Empirical evaluation of gated recurrent neural networks on sequence modeling. https://arxiv.org/abs/1412.3555

[8]Deng LW, Zhao Y, Fu ZD, et al., 2022. Efficient trajectory similarity computation with contrastive learning. Proc 31^st ACM Int Conf on Information & Knowledge Management, p.365-374.

[9]Dhont M, Tsiporkova E, González-Deleito N, 2022. Mining of spatiotemporal trajectory profiles derived from mobility data. IEEE Int Conf on Data Mining Workshops, p.1-9.

[10]Fang ZH, Yang Y, Yang G, et al., 2021. CellSense: human mobility recovery via cellular network data enhancement. Proc ACM Interact Mob Wearab Ubiquit Technol, 5(3):100.

[11]Fang ZQ, Du YT, Zhu XJ, et al., 2022. Spatio-temporal trajectory similarity learning in road networks. Proc 28^th ACM SIGKDD Conf on Knowledge Discovery and Data Mining, p.347-356.

[12]Feng J, Li Y, Zhang C, et al., 2018. DeepMove: predicting human mobility with attentional recurrent networks. Proc World Wide Web Conf, p.1459-1468.

[13]Gao Q, Wang XH, Liu CR, et al., 2023. Open anomalous trajectory recognition via probabilistic metric learning. Proc 32^nd Int Joint Conf on Artificial Intelligence, p.2095-2103.

[14]González MC, Hidalgo CA, Barabási AL, 2008. Understanding individual human mobility patterns. https://arxiv.org/abs/0806.1256

[15]He KM, Fan HQ, Wu YX, et al., 2020. Momentum contrast for unsupervised visual representation learning. Proc IEEE/CVF Conf on Computer Vision and Pattern Recognition, p.9726-9735.

[16]Hristova D, Williams MJ, Musolesi M, et al., 2016. Measuring urban social diversity using interconnected geo-social networks. Proc 25^th Int Conf on World Wide Web, p.21-30.

[17]Li L, Li YB, Li ZH, 2013. Efficient missing data imputing for traffic flow by considering temporal and spatial dependence. Transp Res Part C Emerg Technol, 34:108-120.

[18]Li XC, Zhao KQ, Cong G, et al., 2018. Deep representation learning for trajectory similarity computation. Proc IEEE 34^th Int Conf on Data Engineering, p.617-628.

[19]Li XC, Cong G, Cheng Y, 2020. Spatial transition learning on road networks with deep probabilistic models. Proc IEEE 36^th Int Conf on Data Engineering, p.349-360.

[20]Li YJ, Tarlow D, Brockschmidt M, et al., 2016. Gated graph sequence neural networks. Proc 4^th Int Conf on Learning Representations.

[21]Lin Y, Wan HY, Guo SN, et al., 2021. Pre-training context and time aware location embeddings from spatial-temporal trajectories for user next location prediction. Proc 35^th AAAI Conf on Artificial Intelligence, p.4241-4248.

[22]Liu Q, Wu S, Wang L, et al., 2016. Predicting the next location: a recurrent model with spatial and temporal contexts. Proc 30^th AAAI Conf on Artificial Intelligence, p.194-200.

[23]Luo YH, Cai XR, Zhang Y, et al., 2018. Multivariate time series imputation with generative adversarial networks. Proc 32^nd Int Conf on Neural Information Processing Systems, p.1603-1614.

[24]Noulas A, Shaw B, Lambiotte R, et al., 2015. Topological properties and temporal dynamics of place networks in urban environments. Proc 24^th Int Conf on World Wide Web, p.431-441.

[25]Park D, Kang J, Song H, et al., 2022. Multi-view POI-level cellular trajectory reconstruction for digital contact tracing of infectious diseases. Proc IEEE Int Conf on Data Mining, p.1137-1142.

[26]Ren HM, Ruan SJ, Li YH, et al., 2021. MTrajRec: map-constrained trajectory recovery via Seq2Seq multi-task learning. Proc 27^th ACM SIGKDD Conf on Knowledge Discovery & Data Mining, p.1410-1419.

[27]Salakhutdinov R, Mnih A, 2007. Probabilistic matrix factorization. Proc 20^th Int Conf on Neural Information Processing Systems, p.1257-1264.

[28]Seng D, Lv FS, Liang ZY, et al., 2021. Forecasting traffic flows in irregular regions with multi-graph convolutional network and gated recurrent unit. Front Inform Technol Electron Eng, 22(9):1179-1193.

[29]Si JJ, Yang J, Xiang Y, et al., 2024. TrajBERT: BERT-based trajectory recovery with spatial-temporal refinement for implicit sparse trajectories. IEEE Trans Mob Comput, 23(5):4849-4860.

[30]Sun H, Yang CJ, Deng LW, et al., 2021. PeriodicMove: shift-aware human mobility recovery with graph neural network. Proc 30^th ACM Int Conf on Information & Knowledge Management, p.1734-1743.

[31]Vaswani A, Shazeer N, Parmar N, et al., 2017. Attention is all you need. Proc 31^st Int Conf on Neural Information Processing Systems, p.6000-6010.

[32]Wang YY, Jiang WH, Pu SL, et al., 2020. Learning embeddings of a heterogeneous behavior network for potential behavior prediction. Front Inform Technol Electron Eng, 21(3):422-435.

[33]Wei LY, Zheng Y, Peng WC, 2012. Constructing popular routes from uncertain trajectories. Proc 18^th ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining, p.195-203.

[34]Wu H, Mao JY, Sun WW, et al., 2016. Probabilistic robust route recovery with spatio-temporal dynamics. Proc 22^nd ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining, p.1915-1924.

[35]Wu S, Tang YY, Zhu YQ, et al., 2019. Session-based recommendation with graph neural networks. Proc 33^rd AAAI Conf on Artificial Intelligence, p.346-353.

[36]Xi DB, Zhuang FZ, Liu YC, et al., 2019. Modelling of bi-directional spatio-temporal dependence and users’ dynamic preferences for missing POI check-in identification. Proc 33^rd AAAI Conf on Artificial Intelligence, p.5458-5465.

[37]Xia T, Qi YH, Feng J, et al., 2021. AttnMove: history enhanced trajectory recovery via attentional network. Proc 35^th AAAI Conf on Artificial Intelligence, p.4494-4502.

[38]Xu Y, Xu JJ, Zhao J, et al., 2022. MetaPTP: an adaptive meta-optimized model for personalized spatial trajectory prediction. Proc 28^th ACM SIGKDD Conf on Knowledge Discovery and Data Mining, p.2151-2159.

[39]Yang DQ, Zhang DQ, Zheng VW, et al., 2015. Modeling user activity preference by leveraging user spatial temporal characteristics in LBSNs. IEEE Trans Syst Man Cybern Syst, 45(1):129-142.

[40]Yang S, Liu JM, Zhao KQ, 2022. GETNext: trajectory flow map enhanced Transformer for next POI recommendation. Proc 45^th Int ACM SIGIR Conf on Research and Development in Information Retrieval, p.1144-1153.

[41]Zhang WY, Xia DW, Chang GY, et al., 2022. APFD: an effective approach to taxi route recommendation with mobile trajectory big data. Front Inform Technol Electron Eng, 23(10):1494-1510.

[42]Zhao J, Xu JJ, Zhou R, et al., 2018. On prediction of user destination by sub-trajectory understanding: a deep learning based approach. Proc 27^th ACM Int Conf on Information and Knowledge Management, p.1413-1422.

[43]Zheng Y, Li QN, Chen YK, et al., 2008. Understanding mobility based on GPS data. Proc 10^th Int Conf on Ubiquitous Computing, p.312-321.

[44]Zheng Y, Zhang LZ, Xie X, et al., 2009. Mining interesting locations and travel sequences from GPS trajectories. Proc 18^th Int Conf on World Wide Web, p.791-800.

[45]Zheng Y, Xie X, Ma WY, 2010. GeoLife: a collaborative social networking service among user, location and trajectory. IEEE Data Eng Bull, 33(2):32-39.

[46]Zhou F, Wang PY, Xu X, et al., 2022. Contrastive trajectory learning for tour recommendation. ACM Trans Intell Syst Technol, 13(1):4.