AFGSRec：一种自适应融合全局协同特征的社交推荐模型

doi:10.12141/j.issn.1000-565X.220180

摘要/Abstract

摘要：

以往的序列推荐方法通常从近期交易记录中捕获用户的消费偏好，忽略了全局交易信息和好友偏好对用户交易行为的影响，导致模型的推荐结果不够准确。针对以上问题，文中提出了一种自适应融合全局协同特征的社交推荐模型AFGSRec。首先，用异质图神经网络建模社交网络中的用户、历史交易信息，以捕获全局协同特征和好友之间的社交影响；接着，设计了一种基于选择机制的门图神经网络，以有效过滤与当前序列无关的节点转换信息，更准确地捕获用户当前偏好；然后，提出了一种自适应的特征融合方法，以动态捕获全局协同特征对用户偏好的影响，提高系统的推荐准确率；最后，将周期动态学习率用于模型训练，以更好地处理鞍点，提升模型的收敛速度。实验结果表明：AFGSRec具有较好的鲁棒性，命中率（HR）和平均倒数排名（MRR）都优于当前领先模型SERec，在Gowalla数据集上，HR@10、HR@20分别提升了1.91%和1.15%，MRR@10、MRR@20分别提升了5.05%和4.83%；在Delicious数据集上，HR@10、HR@20分别提升了2.45%和1.19%，MRR@10、MRR@20分别提升了4.84%和4.32%。

Abstract:

Previous session-based recommendation systems usually capture users’ consumption preferences from their recent transaction records, and this method ignores the influence of global transaction information and friends’ preferences on users’ transaction behavior, resulting in less accurate recommendation results of the model. To solve the problem, this paper proposed a social recommendation model AFGSRec based on an adaptive fusion of global collaborative features. Firstly, a heterogeneous graph neural network was used to model users and their historical transaction information on the social network for capturing global collaborative features and social influence among friends. Secondly, this paper designed a graph neural network based on a selection mechanism that effectively filters out the node transition features irrelevant to the current session and captures user preferences more accurately. Thirdly, an adaptive fusion method was designed to capture the impact of global collaborative features on users’ current preferences dynamically and improve the model’s recommendation accuracy. Finally, this paper used a dynamic cyclical learning rate to help the model better handle saddle points during the training process to improve the convergence speed of model AFGSRec. The experimental results show that AFGSRec is robust; both the HR (Hit Rate) and MRR (Mean Reciprocal Rank) of AFGSRec outperform the state-of-art model SERec. On the Gowalla dataset, HR@10 and HR@20 are increased by 1.91% and 1.15%, respectively; MRR@10 and MRR@20 are increased by 5.05% and 4.83%, respectively. On the Delicious dataset, HR@10 and HR@20 are increased by 2.45% and 1.19%, respectively; MRR@10 and MRR@20 are increased by 4.84% and 4.32%, respectively.

Key words: graph neural network, recommendation system, social networking, global collaborative feature, dynamic cyclical learning rate

中图分类号:

TP391

蔡晓东, 曾志杨. AFGSRec：一种自适应融合全局协同特征的社交推荐模型[J]. 华南理工大学学报(自然科学版), 2022, 50(12): 71-79.

CAI Xiaodong, ZENG Zhiyang . AFGSRec: A Social Recommendation Model Based on Adaptive Fusion of Global Collaborative Features[J]. Journal of South China University of Technology(Natural Science Edition), 2022, 50(12): 71-79.

图/表 9

图1

图2

图3

图4

表1

表2

表3

表4

图5

参考文献 25

1	WANG S， HU L， WANG Y，et al ．Sequential recommender systems：challenges，progress and prospects ［C］∥ Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence Survey Track.Macao：Morgan Kaufmann，2019：6332-6338.
2	HIDASI B， KARATZOGLOU A， BALTRUNAS L，et al ．Session-based recommendations with recurrent neural networks ［C］∥ Proceedings of the 4th International Confe-rence on Learning Representations．San Juan：OpenReview.net，2016：1-10.
3	CHO K， MERRIENBOER B V， GULCEHRE C，et al ．Learning phrase representations using RNN encoder-decoder for statistical machine translation ［C］∥ Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing．Doha：Association for Computational Linguistics，2014：1724-1734.
4	BAHDANAU D， CHO K， BENGIO Y ．Neural machine translation by jointly learning to align and translate ［C］∥ Proceedings of the 3rd International Conference on Lear-ning Representations．San Diego：OpenReview.net，2015：1-15.
5	LI J， REN P， CHEN Z，et al ．Neural attentive session-based recommendation ［C］∥ Proceedings of the 2017 ACM Conference on Information and Knowledge Management．Singapore：ACM，2017：1419-1428.
6	REN P， CHEN Z， LI J，et al ．RepeatNet：a repeat aware neural recommendation machine for session-based recommendation ［C］∥ Proceedings of the 33rd AAAI Conference on Artificial Intelligence．Honolulu：AAAI，2019：4806-4813.
7	JI W， WANG K， WANG X，et al ．Sequential recommender via time-aware attentive memory network ［C］∥ Proceedings of the 29th ACM International Conference on Information and Knowledge Management．New York：ACM，2020：565-574.
8	ZHANG J， MA C， MU X，et al ．Recurrent convolutional neural network for session-based recommendation ［J］．Neurocomputing，2021，437：157-167.
9	WU S， TANG Y， ZHU Y，et al ．Session-based recommendation with graph neural networks ［C］∥ Proceedings of the 33rd AAAI Conference on Artificial Intelligence．Honolulu：AAAI，2019：346-353.
10	LI Y， ZEMEL R， BROCKSCHMIDT M，et al ．Gated graph sequence neural networks ［C］∥ Proceedings of the 4th International Conference on Learning Representations．San Juan：OpenReview.net，2016：11-30.
11	WANG Z， WEI W， CONG G，et al ．Global context enhanced graph neural networks for session-based re-commendation ［C］∥ Proceedings of the 43rd International ACM SIGIR Conference on Research and Deve-lopment in Information Retrieval．New York：ACM，2020：169-178.
12	ZHANG C， SONG D， HUANG C，et al ．Heterogeneous graph neural network ［C］∥ Proceedings of the 25th ACM SIGKDD International Conference on Know-ledge Discovery & Data Mining．Anchorage：ACM，2019：793-803.
13	XU L， XI W D， WANG C D ．Session-based recommendation with heterogeneous graph neural networks ［C］∥ Proceedings of 2021 International Joint Confe-rence on Neural Networks．Shenzhen：IEEE，2021：1-8.
14	PANG Y， WU L， SHEN Q，et al ．Heterogeneous global graph neural networks for personalized session-based recommendation ［C］∥ Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining．New York：ACM，2022：775-783.
15	SONG W， XIAO Z， WANG Y，et al ．Session-based social recommendation via dynamic graph attention networks ［C］∥ Proceedings of the Twelfth ACM International Conference on Web Search and Data Mining．Melbourne：ACM，2019：555-563.
16	GU P， HAN Y， GAO W，et al ．Enhancing session-based social recommendation through item graph embedding and contextual friendship modeling［J］．Neurocomputing，2021，419：190-202.
17	CHEN T， WONG R C W ．An efficient and effective framework for session-based social recommendation ［C］∥ Proceedings of the Forteenth ACM International Conference on Web Search and Data Mining．New York：ACM，2021：400-408.
18	MA C， KANG P， LIU X ．Hierarchical gating networks for sequential recommendation ［C］∥ Proceedings of the 25th ACM SIGKDD International Conference on Know-ledge Discovery & Data Mining．Anchorage：ACM，2019：825-833.
19	SMITH L N ．Cyclical learning rates for training neural networks ［C］∥ Proceedings of 2017 IEEE Winter Conference on Applications of Computer Vision．Santa Rosa：IEEE，2017：464-472.
20	WANG C， ZHU Y， LIU H，et al ．Enhancing user interest modeling with knowledge-enriched itemsets for sequential recommendation ［C］∥ Proceedings of the 30th ACM International Conference on Information & Know-ledge Management．New York：ACM，2021：1889-1898.
21	YU F， ZHU Y， LIU Q，et al ．TAGNN：target attentive graph neural networks for session-based recommendation ［C］∥ Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval．New York：ACM，2020：1921-1924.
22	PAN Z， CAI F， CHEN W，et al ．Star graph neural networks for session-based recommendation ［C］∥ Proceedings of the 29th ACM International Conference on Information & Knowledge Management．New York：ACM，2020：1195-1204.
23	CHEN T， WONG R C W ．Handling information loss of graph neural networks for session-based recommendation ［C］∥ Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining．New York：ACM，2020：1172-1180.
24	宋美佳，贾鹤鸣，林志兴，等．自适应学习率梯度下降的优化算法［J］．三明学院学报，2021，38（6）：36-44.
	SONG Meijia， JIA Heming， LIN Zhixing，et al ．Optimization research of gradient descent algorithm based on adaptive learning rate ［J］．Journal of Sanming University，2021，38（6）：36-44.
25	DAUPHIN Y， de VRIES H， BENGIO Y ．Equilibrated adaptive learning rates for non-convex optimization ［C］∥ Proceedings of the 28th International Conference on Neural Information Processing Systems．Cambridge：MIT Press，2015：1504-1512.

数据集	用户数	交易物品数	交易次数	社交关系	序列数	用户平均邻居节点数	用户平均交易数	单个序列平均交易数
Gowalla	33 661	41 229	1 218 599	283 778	258 732	8.43	36.20	4.71
Delicious	1 313	5 793	266 190	9 130	60 397	6.95	202.73	4.41

K	模型	Gowalla		Delicious
K	模型	HR/%	MRR/%	HR/%	MRR/%
10	SR-GNN	41.31	22.39	37.01	19.57
	DGRec	42.18	23.04	37.78	20.07
	SERec	46.01	25.14	40.02	21.29
	AFGSRec	46.89	26.41	41.00	22.32
20	SR-GNN	49.31	22.94	45.74	20.20
	DGRec	49.95	23.58	47.36	20.73
	SERec	53.72	25.67	49.53	21.98
	AFGSRec	54.34	26.91	50.12	22.93

K	模型	Gowalla		Delicious
K	模型	HR/%	MRR/%	HR/%	MRR/%
10	AFGSRec-wSPI	46.31	26.02	40.49	22.09
	AFGSRec-wAFC	46.43	26.18	40.28	21.91
	AFGSRec-wCYC	46.22	25.39	40.58	22.10
	AFGSRec	46.89	26.41	41.00	22.32
20	AFGSRec-wSPI	53.83	26.03	49.70	22.52
	AFGSRec-wAFC	53.74	26.33	49.71	22.57
	AFGSRec-wCYC	53.55	25.87	49.81	22.60
	AFGSRec	54.34	26.91	50.12	22.93

K	融合方式	Gowalla		Delicious
K	融合方式	HR/%	MRR/%	HR/%	MRR/%
10	CAT	45.96	25.21	40.21	21.65
	SUM	45.60	25.14	40.37	21.57
	AFC	46.19	25.48	40.46	21.70
20	CAT	53.69	26.03	49.54	22.27
	SUM	53.02	25.64	49.60	22.18
	AFC	53.95	26.00	49.62	22.38

[1]	蔡晓东, 周青松, 叶青. 基于动态邻域采样的社交推荐模型[J]. 华南理工大学学报(自然科学版), 2024, 52(2): 32-41.
[2]	赵荣超, 吴百礼, 陈祝云, 温楷儒, 张绍辉, 李巍华. 多尺度时空信息融合驱动的图神经网络故障诊断方法[J]. 华南理工大学学报(自然科学版), 2023, 51(12): 42-52.
[3]	许伦辉, 余佳芯, 裴明阳, 等. 基于几何路网结构和强化学习的车辆重定位策略[J]. 华南理工大学学报(自然科学版), 2023, 51(10): 99-109.
[4]	陆以勤, 潘周双, 张洋, 等. 基于交互式连接图注意力网络的知识图谱补全方法[J]. 华南理工大学学报(自然科学版), 2022, 50(12): 13-19.
[5]	蔡晓东洪涛曹艺. 基于极化关系表述与低维数据间关联学习的推荐模型[J]. 华南理工大学学报（自然科学版）, 2022, 50(1): 122-131.
[6]	黄敏齐海涛蒋春林. 基于注意力机制的耦合协同过滤模型[J]. 华南理工大学学报(自然科学版), 2021, 49(7): 59-65.
[7]	古万荣, 谢贤芬, 张子烨, 等. 面向非随机缺失数据的协同过滤评分方法[J]. 华南理工大学学报（自然科学版）, 2021, 49(1): 47-57.
[8]	刘慧婷, 纪强, 刘慧敏, 等. 基于双层注意力机制的联合深度推荐模型[J]. 华南理工大学学报（自然科学版）, 2020, 48(6): 97-105.
[9]	王伟凝, 马学东, 苏俊杰, 等. 基于个性化生成模板集聚的用户图像美感表征[J]. 华南理工大学学报（自然科学版）, 2020, 48(3): 1-9.
[10]	张子烨, 李明畅, 梁凌睿, 等. 推荐系统信息跨领域的改进迁移学习算法[J]. 华南理工大学学报（自然科学版）, 2020, 48(11): 99-106.
[11]	李家春熊冬青曹建洲. 一种移动社交网络的协同定位隐私保护方法 [J]. 华南理工大学学报（自然科学版）, 2019, 47(2): 92-97,105.
[12]	杜卿王齐轩黄东平蔡毅王涛闵华清. 基于社交关系的问答系统及最佳回答者推荐技术[J]. 华南理工大学学报（自然科学版）, 2015, 43(1): 132-139.