结合通用轨迹图和多偏好的兴趣点推荐方法

doi:10.3969/j.issn.0255-8297.2025.05.005

摘要/Abstract

摘要： 传统兴趣点（point of interest,POI）推荐方法对用户和POI的关联关系挖掘不充分，无法全面捕捉用户偏好；基于图增强的推荐方法虽能挖掘关联关系，却易引入噪声，降低推荐性能。针对这些问题，本文提出了结合通用轨迹图和多偏好的POI推荐方法。首先构建了用户与POI的带权二部图，利用图卷积网络捕捉用户和POI的交互关系，学习用户兴趣偏好；利用兴趣偏好完成用户聚类，进而构建同类型用户通用轨迹图，减少噪声信息影响；利用图卷积网络捕捉同类型用户的群体特征，丰富特征表示。其次，将群体特征与用户当前轨迹中时间类别感知信息、时空上下文信息相结合，利用Transformer挖掘用户的深层行为偏好。再次，构造非线性加性函数并将兴趣偏好和行为偏好动态组合，全面捕捉用户偏好，完成POI推荐。最后，在真实数据集上验证了本文方法的有效性。

关键词: 兴趣点, 用户偏好, 推荐系统, 图卷积网络, 通用轨迹图

Abstract: Traditional point of interest (POI) recommendation methods often fail to fully mine the relationships between users and POIs, resulting in limited capacity to capture user preferences. Although graph enhancement method offers improved relational modeling, it is susceptible to noise, which reduces the recommendation precision. To address these challenges, this paper proposes a POI recommendation method named combining universal trajectory maps and multiple preferences. Firstly, a weighted bipartite graph between users and POIs is constructed. Graph convolutional network (GCN) is used to extract the interactive relationship between users and POIs to learn users’ interest preferences. Users are clustered using the obtained interest preferences. The general trajectory maps of the same type of users are built to reduce the impact of noise information. GCN is further used to mine the group features of different types of users and enrich feature representation. Secondly, the group features are combined with time-aware category information and spatio-temporal context from the user’s current trajectory. Transformer model is used to capture deep behavioral preferences. Finally, a non-linear additive function is used to dynamically combine interest preferences with current behavior preferences to fully capture user preferences and generate POI recommendations. The validity of the proposed method is verified on real data sets.

Key words: point of interest (POI), user preferences, recommendation system, graph convolutional network (GCN), universal trajectory map

中图分类号:

P751.1

卢菁, 葛聪. 结合通用轨迹图和多偏好的兴趣点推荐方法[J]. 应用科学学报, 2025, 43(5): 771-784.

LU Jing, GE Cong. Point of Interest Recommendation Method Combining Universal Trajectory Maps and Multiple Preferences[J]. Journal of Applied Sciences, 2025, 43(5): 771-784.

参考文献

[1] 任星怡, 宋美娜, 宋俊德. 基于位置社交网络的上下文感知的兴趣点推荐[J]. 计算机学报, 2017, 40(4): 824-841. Ren X Y, Song M N, Song J D. Context-aware point-of-interest recommendation in locationbased social networks [J]. Chinese Journal of Computers, 2017, 40(4): 824-841. (in Chinese)
[2] Lian D F, Ge Y, Zhang F Z, et al. Content-aware collaborative filtering for location recommendation based on human mobility data [C]//2015 IEEE International Conference on Data Mining, 2015: 261-270.
[3] Rendle S, Freudenthaler C, Schmidt-Thieme L. Factorizing personalized Markov chains for next-basket recommendation [C]//19th International Conference on World Wide Web, 2010: 811-820.
[4] Xie M, Yin H Z, Wang H, et al. Learning graph-based POI embedding for location-based recommendation [C]//25th ACM International on Conference on Information and Knowledge Management, 2016: 15-24.
[5] Wang H W, Zhao M, Xie X, et al. Knowledge graph convolutional networks for recommender systems [C]//The World Wide Web Conference, 2019: 3307-3313.
[6] Halilaj L, Lüttin J, Rothermel S, et al. Towards a knowledge graph-based approach for context-aware points-of-interest recommendations [C]//36th Annual ACM Symposium on Applied Computing, 2021: 1846-1854.
[7] Liu Q, Wu S, Wang L, et al. Predicting the next location: a recurrent model with spatial and temporal contexts [C]//Thirtieth AAAI Conference on Artificial Intelligence, 2016:194-200.
[8] Wu Y X, Li K, Zhao G S, et al. Long- and short-term preference learning for next POI recommendation [C]/ 28th ACM International Conference on Information and Knowledge Management, 2019: 2301-2304.
[9] Liu Y, Zhu H P, Chen Y, et al. Long- and short-term preference model based on graph embedding for sequential recommendation [C]//DASFAA 2020 International Workshops, 2020: 241-257.
[10] Zhao P P, Luo A J, Liu Y C, et al. Where to go next: a spatio-temporal gated network for next POI recommendation [C]//IEEE Transactions on Knowledge and Data Engineering, 2022: 2512-2524.
[11] Cui Q, Tang Y Y, Wu S, et al. Distance2Pre: personalized spatial preference for next pointof-interest prediction [C]//Advances in Knowledge Discovery and Data Mining, 2019: 289-301.
[12] Lian D F, Wu Y J, Ge Y, et al. Geography-aware sequential location recommendation [C]//26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2020: 2009-2019.
[13] Luo Y T, Liu Q, Liu Z C. STAN: spatio-temporal attention network for next location recommendation [C]//The World Wide Web Conference, 2021: 2177-2185.
[14] Yang S, Liu J M, Zhao K Q. GETNext: trajectory flow map enhanced transformer for next POI recommendation [C]//45th International ACM SIGIR Conference on Research and Development in Information Retrieval, 2022: 1144-1153.
[15] Wu Y X, Li K, Zhao G S, et al. Long- and short-term preference learning for next POI recommendation [C]//28th ACM International Conference on Information and Knowledge Management, 2019: 2301-2304.
[16] Sun K, Qian T Y, Chen T, et al. Where to go next: modeling long- and short-term user preferences for point-of-interest recommendation [J]. AAAI Conference on Artificial Intelligence, 2020, 34(1): 214-221.
[17] Wu Y X, Li K, Zhao G S, et al. Personalized long- and short-term preference learning for next POI recommendation [J]. IEEE Transactions on Knowledge and Data Engineering, 2022, 34(4): 1944-1957.
[18] Wang X, He X N, Wang M, et al. Neural graph collaborative filtering [C]//42nd International ACM SIGIR Conference on Research and Development in Information Retrieval, 2019: 165-174.
[19] He X N, Deng K, Wang X, et al. LightGCN: simplifying and powering graph convolution network for recommendation [C]//43rd International ACM SIGIR Conference on Research and Development in Information Retrieval, 2020: 639-648.
[20] Zahra S, Ali Ghazanfar M, Khalid A, et al. Novel centroid selection approaches for KMeans-clustering based recommender systems [J]. Information Sciences, 2015, 320: 156-189.
[21] Peñaloza V. Time2Vec embedding on a Seq2Seq Bi-directional LSTM network for pedestrian trajectory prediction [J]. Research on Computing Science, 2020, 149(11): 249-260.
[22] Tobler W R. A computer movie simulating urban growth in the Detroit Region [J]. Economic Geography, 1970, 46: 234.
[23] Xie J Y, Chen Z Z. Hierarchical transformer with spatio-temporal context aggregation for next point-of-interest recommendation [J]. ACM Transactions on Information Systems, 2024, 42(2): 1-30.
[24] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need [C]//Conference and Workshop on Neural Information Processing Systems, 2017: 5998-6008.