基于多因素解纠缠的用户—兴趣点联合预测

doi:10.3969/j.issn.1671-1122.2024.11.008

信息网络安全 ›› 2024, Vol. 24 ›› Issue (11): 1685-1695.doi: 10.3969/j.issn.1671-1122.2024.11.008

基于多因素解纠缠的用户—兴趣点联合预测

马卓¹, 陈东子²(), 何佳涵¹, 王群¹

1.江苏警官学院计算机信息与网络安全系，南京 210031
2.南京市公安局，南京 210005

收稿日期:2024-07-11 出版日期:2024-11-10 发布日期:2024-11-21
通讯作者: 陈东子 chendz321@hotmail.com
作者简介:马卓（1993—），女，山西，讲师，博士，CCF会员，主要研究方向为信息安全、用户隐私|陈东子（1993—），男，河南，硕士，主要研究方向为网络空间安全|何佳涵（2002—），女，江苏，本科，主要研究方向为网络空间安全|王群（1971—），男，甘肃，教授，博士，CCF杰出会员，主要研究方向为网络空间安全
基金资助:
国家自然科学基金(62202209)

Joint Prediction for User and Point of Interest Based on Disentangling Influences

MA Zhuo¹, CHEN Dongzi²(), HE Jiahan¹, WANG Qun¹

1. Department of Computer Information and Cybersecurity, Jiangsu Police Institute, Nanjing 210031, China
2. Nanjing Municipal Public Security Bureau, Nanjing 210005, China

Received:2024-07-11 Online:2024-11-10 Published:2024-11-21

摘要/Abstract

摘要：

用户—兴趣点预测问题基于在线用户的历史签入记录来判断用户是否会签入特定兴趣点，但用户—兴趣点数据存在长尾分布现象。针对该数据稀疏性问题，一些研究人员将地理效应和地理序列效应通过自监督学习进行解纠缠表示，以提升兴趣点预测任务的可解释性和准确性。文章引入语义序列效应，提出一种改进的解纠缠图嵌入模型，该模型利用兴趣点在地理空间和语义空间的成对约束，基于地理坐标空间和语义类别空间中影响因素的特征表达、特征修正、特征解耦合和多层感知机融合，在地理层面上结合语义层面更好地预测用户对兴趣点的访问情况。实验结果表明，该方法在签入稀疏的数据集上依然能够取得良好的预测效果。

关键词: 兴趣点预测, 自监督解纠缠, 图嵌入, 联合预测

Abstract:

The problem of user-POI prediction, based on the user’s historical check-in records, determines whether a user checks in a specific POI. However, the user-POI data has a long-tail distribution phenomenon. To address this data sparsity challenge, existing work disentangled the geographical neighbor effect and the geographical sequence effect through self-supervised learning to improve the interpretability and accuracy of the POI prediction task. This paper further introduced the semantic sequence effect, and proposed an improved disentangled graph embedding model. The model used the pairwise constraints of point-of-interests in the geographic space and semantic space, and was based on the feature expression, feature modification, feature decoupling and multi-layer perceptron fusion of the influencing factors in the geographic coordinate space and the semantic category space. The geographic level could be combined with the semantic level to better predict the user’s access to the POI. Experimental results show that the proposed method can still achieve good prediction effects on sparse datasets.

Key words: point of interest prediction, self-supervised disentanglement, graph embedding, joint prediction

中图分类号:

TP309

马卓, 陈东子, 何佳涵, 王群. 基于多因素解纠缠的用户—兴趣点联合预测[J]. 信息网络安全, 2024, 24(11): 1685-1695.

MA Zhuo, CHEN Dongzi, HE Jiahan, WANG Qun. Joint Prediction for User and Point of Interest Based on Disentangling Influences[J]. Netinfo Security, 2024, 24(11): 1685-1695.

图/表 6

图1

图2

表1

图3

图4

图5

参考文献 18

[1]	CHO E, MYERS S A, LESKOVEC J. Friendship and Mobility: User Movement in Location-Based Social Networks[C]//ACM. The 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2011: 1082-1090.
[2]	QIN Yifang, WANG Yifan, SUN Fang, et al. DisenPOI: Disentangling Sequential and Geographical Influence for Point-of-Interest Recommendation[C]//ACM. The 6th ACM International Conference on Web Search and Data Mining. New York: ACM, 2023: 508-516.
[3]	HE Xiangnan, DENG Kuan, WANG Xiang, et al. LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation[C]//ACM. The 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval 2020. New York: ACM, 2020: 639-648.
[4]	SUN Ke, QIAN Tieyun, CHEN Tong, et al. Where to Go Next: Modeling Long- and Short-Term User Preferences for Point-of-Interest Recommendation[C]//ACM. The AAAI Conference on Artificial Intelligence. New York: ACM, 2020: 214-221.
[5]	WANG Zhaobo, ZHU Yanmin, ZHANG Qiaomei, et al. Graph-Enhanced Spatial-Temporal Network for Next POI Recommendation[J]. ACM Transactions on Knowledge Discovery from Data, 2022, 16(6): 1-21.
[6]	XIAO Lei, LI Qi. Survey of Temporal Knowledge Graph Completion Methods[J]. Computer Engineering and Applications, 2024, 60(6): 43-54. doi: 10.3778/j.issn.1002-8331.2307-0083
	肖蕾, 李琪. 时序知识图谱补全方法研究综述[J]. 计算机工程与应用, 2024, 60(6):43-54. doi: 10.3778/j.issn.1002-8331.2307-0083
[7]	ZHAO Jing, XU Jiajie, XU Yuan, et al. CCML: Curriculum and Contrastive Learning Enhanced Meta-Learner for Personalized Spatial Trajectory Prediction[J]. IEEE Transactions on Knowledge and Data Engineering, 2024, 36(9): 4499-4514.
[8]	ZHANG Qingbo, WANG Bin, CUI Ningning, et al. Attention-Based Regularized Matrix Factorization for Recommendation[J]. Journal of Software, 2020, 31(3): 778-793.
	张青博, 王斌, 崔宁宁, 等. 基于注意力机制的规范化矩阵分解推荐算法[J]. 软件学报, 2020, 31(3):778-793.
[9]	PAN Yiteng, HE Fazhi, YU Haiping. Learning Social Representations with Deep Autoencoder for Recommender System[J]. World Wide Web, 2020, 23(4): 2259-2279.
[10]	WEN Tao, ZHANG Min, WANG Huaiyuan. Transient Stability Assessment Model Based on Stacked Sparse Denoising Auto-Encodern[J]. Electric Power Engineering Technology, 2022, 41(1): 207-212.
	温涛, 张敏, 王怀远. 基于堆叠稀疏降噪自编码器的暂态稳定评估模型[J]. 电力工程技术, 2022, 41(1):207-212.
[11]	ZHENG Yu, GAO Chen, CHANG Jianxin, et al. Disentangling Long and Short-Term Interests for Recommendation[C]//ACM. Proceedings of the ACM Web Conference 2022. New York: ACM, 2022: 2256-2267.
[12]	LIU Guoqi, HE Tingnian, RONG Yixuan, et al. A Point of Interest Recommendation Model Based on Tracks and Friend Relationship of Users[EB/OL]. (2024-03-22)[2024-06-20]. https://kns.cnki.net/kcms2/article/abstract?v=tH39KOVtnoGMPmh6qUh_IX6qyKDlvcDYhTSYDyCq40SnG8jamxDzNr59fhLWxIwal1stTMjKiE3t8ubpU-yAJ7K_EIEKa8UeDECtDEXG1Oe_a8Aw4sL3foaAh5xBND34G1375xYA3rZPzhszd2knpUsKLpNzVtahX__tQkTPWtBi0wUsJ0b371PDwkoIanD_&uniplatform=NZKPT&language=CHS.
	刘国岐, 何廷年, 荣艺煊, 等. 基于用户轨迹和好友关系的兴趣点推荐[EB/OL]. (2024-03-22)[2024-06-20]. https://kns.cnki.net/kcms2/article/abstract?v=tH39KOVtnoGMPmh6qUh_IX6qyKDlvcDYhTSYDyCq40SnG8jamxDzNr59fhLWxIwal1stTMjKiE3t8ubpU-yAJ7K_EIEKa8UeDECtDEXG1Oe_a8Aw4sL3foaAh5xBND34G1375xYA3rZPzhszd2knpUsKLpNzVtahX__tQkTPWtBi0wUsJ0b371PDwkoIanD_&uniplatform=NZKPT&language=CHS.
[13]	CAI Jiawei, WANG Dong, CHEN Hongyang, et al. Modeling Dynamic Spatiotemporal User Preference for Location Prediction: A Mutually Enhanced Method[J]. World Wide Web, 2024, 27(2): 1-17.
[14]	ZHOU Guorui, MOU Na, FAN Ying, et al. Deep Interest Evolution Network for Click-through Rate Prediction[C]// ACM. The AAAI Conference on Artificial Intelligence. New York: ACM, 2019: 5941-5948.
[15]	ZHOU Guorui, ZHU Xiaoqiang, SONG Chenru, et al. Deep Interest Network for Click-through Rate Prediction[C]// ACM. The 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York: ACM, 2018: 1059-1068.
[16]	WU Shu, TANG Yuyuan, ZHU Yanqiao, et al. Session-Based Recommendation with Graph Neural Networks[C]// ACM. The AAAI Conference on Artificial Intelligence. New York: ACM, 2019: 346-353.
[17]	WANG Xiang, HE Xiangnan, WANG Meng, et al. Neural Graph Collaborative Filtering[C]// ACM. The 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: ACM, 2019: 165-174.
[18]	WANG Hao, SHEN Huawei, OUYANG Wentao, et al. Exploiting POI-Specific Geographical Influence for Point-of-Interest Recommendation[C]// ACM. The 27th International Joint Conference on Artificial Intelligence. New York: ACM, 2018: 3877-3883.

模型	纽约数据集		东京数据集
模型	AUC	Logloss	AUC	Logloss
DIN	0.8015	0.5203	0.8623	0.4174
DIEN	0.8062	0.5582	0.8643	0.4544
SR-GNN	0.8491	0.5262	0.8867	0.4481
NGCF	0.8461	0.4788	0.8899	0.4436
LightGCN	0.8729	0.4815	0.8998	0.4348
GeoIE	0.8769	0.4810	0.9083	0.4203
LSTPM	0.8568	0.4821	0.8745	0.4374
GSTN	0.8521	0.5323	0.8906	0.4312
DisenPOI	0.9030	0.4181	0.9364	0.3463
mDisenPOI	0.9155	0.3824	0.9503	0.2900

基于多因素解纠缠的用户—兴趣点联合预测

Joint Prediction for User and Point of Interest Based on Disentangling Influences

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 6

参考文献 18

相关文章 1

编辑推荐

Metrics

本文评价