Home Big Data Mining and Analytics Article
PDF (1.6 MB)
Cite
EndNote(RIS) BibTeX
Collect
Submit Manuscript
Show Outline
Figures (8)

Tables (4)
Table 1
Table 2
Table 3
Table 4
Open Access

IDBR: Interaction-Aware Dual-Granularity Learning for Bundle Recommendation

School of Computer Science (National Pilot Software Engineering School), Beijing University of Posts and Telecommunications, Beijing 100876, China, and also with Xiangjiang Laboratory, Changsha 410205, China
Faculty of Science, National University of Singapore, Singapore 119077, Singapore
School of Computer Science (National Pilot Software Engineering School) and Key Laboratory of Trustworthy Distributed Computing and Service of Ministry of Education, Beijing University of Posts and Telecommunications, Beijing 100876, China
College of Information Science and Technology, Jinan University, Guangzhou 510632, China
Xiangjiang Laboratory, Changsha 410205, China, and also with Business School, Central South University, Changsha 410083, China
Show Author Information

Abstract

In the recommendation system, bundle recommendation is a prevalent sales strategy in which a combination of diverse, related, or complementary products is suggested to consumers. Recent methodologies frequently utilize graph neural networks to capture information from user-bundle, user-item, and bundle-item interactions, deriving corresponding feature representations. However, these approaches often emphasize the distinctions among these three interaction types or treat them uniformly, neglecting the varying importance within one type of interaction and failing to consider the acquisition of information at varying granularities from different types of interactions. In this study, we employ a graph attention mechanism to process user-bundle interaction information, and optimize it using an association enhancement method to extract and construct coarse-grained information representations for users and bundles. By analyzing interactions between users and items, as well as between bundles and items, we identify disparities in item popularity and update the items’ feature representations, facilitating the acquisition of fine-grained information representations for users and bundles. By merging this information, we achieve more comprehensive representations of user intent and bundle characteristics. Extensive experiments on two real-world datasets convincingly demonstrate that our approach significantly advances the task of bundle recommendation, outperforming state-of-the-art methods.

Keywords

recommendation systembundle recommendationgraph attention mechanismcoarse-grained informationfine-grained information

References

[1]

M. Niranjanamurthy, N. Kavyashree, S. Jagannath, and D. Chahar, Analysis of e-commerce and m-commerce: Advantages, limitations and security issues, Int. J. Adv. Res. Comput. Commun. Eng., vol. 2, no. 6, pp. 2360–2370, 2013.

Google Scholar
[2]

Y. Heng, Z. Gao, Y. Jiang, and X. Chen, Exploring hidden factors behind online food shopping from amazon reviews: A topic mining approach, J. Retail. Consum. Serv., vol. 42, pp. 161–168, 2018.

Crossref Google Scholar
[3]

M. D. Johnson, A. Herrmann, and H. H. Bauer, The effects of price bundling on consumer evaluations of product offerings, Int. J. Res. Mark., vol. 16, no. 2, pp. 129–142, 1999.

Crossref Google Scholar
[4]

R. J. Kauffman and E. A. Walden, Economics and electronic commerce: Survey and directions for research, Int. J. Electron. Commerce, vol. 5, no. 4, pp. 5–116, 2001.

Crossref Google Scholar
[5]
Q. Deng, K. Wang, M. Zhao, Z. Zou, R. Wu, J. Tao, C. Fan, and L. Chen, Personalized bundle recommendation in online games, in Proc. 29th ACM Int. Conf. Inf. Knowl. Manag., New York, NY, USA, 2020, pp. 2381–2388.
Crossref
[6]
J. Chang, C. Gao, X. He, D. Jin, and Y. Li, Bundle recommendation with graph convolutional networks, in Proc. 43rd Int. ACM SIGIR Conf. Res. Develop. Inf. Retr., New York, NY, USA, 2020, pp. 1673–1676.
Crossref
[7]
Y. Ma, Y. He, A. Zhang, X. Wang, and T.-S. Chua, Crosscbr: Cross-view contrastive learning for bundle recommendation, in Proc. 28th ACM SIGKDD Conf. Knowl. Discov. Data Min., New York, NY, USA, 2022, pp. 1233–1241.
Crossref
[8]
X. He, K. Deng, X. Wang, Y. Li, Y. Zhang, and M. Wang, Lightgcn: Simplifying and powering graph convolution network for recommendation, in Proc. 43rd Int. ACM SIGIR Conf. Res. Develop. Inf. Retr., New York, NY, USA, 2020, pp. 639–648.
Crossref
[9]
T. N. Kipf and M. Welling.Semi-supervised classification with graph convolutional networks, in Proc. 5th Int. Conf. Learn. Represent., Toulon, France, 2017.
[10]
X. Wang, X. He, M. Wang, F. Feng, and T.-S. Chua, Neural graph collaborative filtering, in Proc. 42nd Int. ACM SIGIR Conf. Res. Develop. Inf. Retr., New York, NY, USA, 2019, pp. 165–174.
Crossref
[11]
S. Zhao, W. Wei, D. Zou, and X. Mao, Multi-view intent disentangle graph networks for bundle recommendation, in Proc. AAAI Conf. Artif. Intell., 2022, vol. 36, no. 4, pp. 4379–4387.
Crossref
[12]
Z.-Y. Zhao, C.-D. Wang, P.-J. Zheng, Q. Gong, K.-W. Huang, and J.-H. Lai, Music sharing platform based on sina app engine, in Proc. 9th Int. Conf. Front. Comput. Sci. Technol., Dalian, China, 2015, pp. 298–303.
Crossref
[13]

R. E. Nisbett and T. D. Wilson, The halo effect: Evidence for unconscious alteration of judgments, J. Pers. Soc. Psychol., vol. 35, no. 4, pp. 250–256, 1977.

Crossref Google Scholar
[14]

M. Haenlein, B. Libai, and E. Muller, Satiation and cross promotion: Selling and swapping users in mobile games, Int. J. Res. Mark., vol. 40, no. 2, pp. 342–361, 2023.

Crossref Google Scholar
[15]

M. Beladev, L. Rokach, and B. Shapira, Recommender systems for product bundling, Knowl.-Based Syst., vol. 111, pp. 193–206, 2016.

Crossref Google Scholar
[16]
A. Pathak, K. Gupta, and J. McAuley, Generating and personalizing bundle recommendations on steam, in Proc. 40th Int. ACM SIGIR Conf. Res. Develop. Inf. Retr., New York, NY, USA, 2017, pp. 1073–1076.
Crossref
[17]
L. Chen, Y. Liu, X. He, L. Gao, and Z. Zheng, Matching user with item set: Collaborative bundle recommendation with deep attention network, in Proc. 28th Int. Joint Conf. Artif. Intell.,, Macau, China, 2019, pp. 2095–2101.
Crossref
[18]
D. Cao, L. Nie, X. He, X. Wei, S. Zhu, and T.- S. Chua, Embedding factorization models for jointly recommending items and user generated lists, in Proc. 40th Int. ACM SIGIR Conf. Res. Develop. Inf. Retr., New York, NY, USA, 2017, pp. 585–594.
Crossref
[19]
X. Wang, X. Liu, J. Liu, and H. Wu, Relational graph neural network with neighbor interactions for bundle recommendation service, in Proc. IEEE Int. Conf. Web Serv., Chicago, IL, USA, 2021, pp. 167–172.
Crossref
[20]
M. Tan, W. Chen, W. Wang, A. Liu, and L. Zhao, Intention-oriented hierarchical bundle recommendation with preference transfer, in Proc. IEEE Int. Conf. Web Serv., Chicago, IL, USA, 2021, pp. 107–116.
Crossref
[21]
P. Velickovic, G. Cucurull, A. Casanova, A. Romero, P. Lio, and Y. Bengio, Graph attention networks, presented at Int. Conf. Learn. Represent., Vancouver, Canada, 2018.
[22]
W. L. Hamilton, R. Ying, and J. Leskovec, Inductive representation learning on large graphs, in Proc. 31st Int. Conf. Neural Inf. Process. Syst., Long Beach, CA, USA, 2017, pp.1025−1035.
[23]
X. Wang, H. Ji, C. Shi, B. Wang, Y. Ye, P. Cui, and P. S. Yu, Heterogeneous graph attention network, in Proc. World Wide Web Conf., New York, NY, USA, 2019, pp. 2022–2032.
Crossref
[24]
S. Abu-El-Haija, B. Perozzi, R. Al-Rfou, and A. Alemi, Watch your step: Learning node embeddings via graph attention, in Proc. 32nd Int. Conf. Neural Inf. Process. Syst., Red Hook, NY, USA, 2018, pp. 9198–9208.
[25]
B. Hu, C. Shi, W. X. Zhao, and P. S. Yu, Leveraging meta path based context for top-n recommendation with a neural co-attention model, in Proc. 24th ACM SIGKDD Int. Conf. Knowl. Discov. Data Min., Barcelona, Spain, 2018, pp. 1531–1540.
Crossref
[26]
A. L. Maas, A. Y. Hannun, and A. Y. Ng, Rectifier nonlinearities improve neural network acoustic models, in Proc. Int. Conf. Mach. Learn., Atlanta, GA, USA, 2013, pp. 1–3.
[27]

R. Bro and A. K. Smilde, Principal component analysis, Anal. Methods, vol. 6, no. 9, pp. 2812–2831, 2014.

Crossref Google Scholar
[28]

A. L. Custodio, H. Rocha, and L. N. Vicente, Incorporating minimum frobenius norm models in direct search, Comput. Optim. Appl., vol. 46, no. 2, pp. 265–278, 2010.

Crossref Google Scholar
[29]

L. Falschlunger, O. Lehner, and H. Treiblmaier, Infovis: The impact of information overload on decision making outcome in high complexity settings, J. Decis. Syst., vol. 25, no. 1, pp. 1–12, 2016.

Crossref Google Scholar
[30]

M. Liu, Z. Tu, H. Xu, X. Xu, and Z. Wang, DysR: A dynamic graph neural network based service bundle recommendation model for mashup creation, IEEE Trans. Serv. Comput., vol. 16, no. 4, pp. 2592–2605, 2023.

Crossref Google Scholar
[31]
J. Zhuo, Q. Zhu, Y. Yue, and Y. Zhao, Learning explicit user interest boundary for recommendation, in Proc. ACM Web Conf., New York, NY, USA, 2022, pp. 193–202.
Crossref
[32]
Loshchilov I, Hutter F. Fixing weight decay regularization in Adam, arXiv preprint arXiv: 1711.05101, 2017.
[33]
S. Rendle, C. Freudenthaler, Z. Gantner, and L. Schmidt Thieme, BPR: Bayesian personalized ranking from implicit feedback, in Proc. 25th Conf. Uncertain. Artif. Intell., Arlington, VA, USA, 2009, pp. 452–461.
[34]
M. Schlichtkrull, T. N. Kipf, P. Bloem, R. van den Berg, I. Titov, and M. Welling, Modeling relational data with graph convolutional networks, in Proc. 15th Int. Semantic Web Conf., Heraklion, Greece, 2018, pp. 593–607.
Crossref
[35]
Z. Liu, M. Wan, S. Guo, K. Achan, and P. S. Yu, BasConv: Aggregating heterogeneous interactions for basket recommendation with graph convolutional neural network, in Proc. SIAM Int. Conf. Data Min., Cincinnati, OH, USA, 2019, pp. 64–72.
Crossref
[36]
D. Chen, Y. Lin, W. Li, P. Li, J. Zhou, and X. Sun. Measuring and relieving the over-smoothing problem for graph neural networks from the topological view, Proc. AAAI Conf. Artif. Intell., vol. 34, no. 4, pp. 3438−3445, 2020.
Crossref
[37]
K. He, X. Zhang, S. Ren, and J. Sun, Delving deep into rectifiers: Surpassing human-level performance on imagenet classification, in Proc. IEEE Int. Conf. Comput. Vis., Santiago, Chile, 2015, pp. 1026–1034.
Crossref
[38]
X. Glorot and Y. Bengio, Understanding the difficulty of training deep feedforward neural networks, in Proc. 13th Int. Conf. Artif. Intell. Stat., Sardinia, Italy, 2010, pp. 249–256.
Big Data Mining and Analytics
Volume 8 Issue 3,
May 2025
Pages 751-766
DOI: 10.26599/BDMA.2025.9020016
Cite this article:
Wang J, Cao Y, Zhang F, et al. IDBR: Interaction-Aware Dual-Granularity Learning for Bundle Recommendation. Big Data Mining and Analytics, 2025, 8(3): 751-766. https://doi.org/10.26599/BDMA.2025.9020016
Metrics & Citations  
Article History
Copyright
Rights and Permissions
Return

Important notation list.

VariableDefinition
eFeature vector of a single entity
EFeature embedding matrix of a certain type of entity
WWeight matrix used in GAT
RU-B interaction matrix with virtual edges
FrobFrobenius norm
ZpbB-I interaction matrix that considers item popularity
MMatrix with co-occurrence relationships
EcFeature embedding matrix in coarse-grained view
EfFeature embedding matrix in fine-grained view
OUsers’ personalized interest boundaries
||Θ22||Regularization term

Statistics of two utilized datasets.

DatasetNumber of usersNumber of bundlesNumber of itemsNumber of U-B interactionsNumber of B-I interactionsNumber of U-I interactionsU-Bdensity (%)B-I density (%)U-I density (%)
NetEase18 52822 864123 628302 3031 778 8381 128 0650.070.060.05
YouShu8039477132 77051 337176 667138 5150.130.110.05

Overall performance comparison.

MethodNetEaseYouShu
R@20N@20R@40N@40R@80N@80R@20N@20R@40N@40R@80N@80
Note: Underlined data represent the best results in the baseline methods, bolded data represent the best results including IDBR. “R” represents “Recall” and “N” represents “NDCG”. The improved ratio (%Improv.) is calculated by comparing the optimal value (bolded) with the suboptimal value (underlined).
MFBPR0.03350.01810.06000.02460.09480.03230.19590.11170.27350.13200.37100.1543
DAM0.04110.02100.06900.02810.10900.03720.20820.11980.28900.14180.39150.1658
RGCN0.04070.02100.06700.02800.11120.03780.20400.10690.30170.13300.41690.1595
BaseConv0.04700.02440.07860.03270.12410.04290.21210.12040.29780.14340.39880.1548
BGCN0.04910.02580.08290.03460.13040.04530.23470.13450.32480.15930.43550.1851
MIDGN0.06780.03430.10850.04510.16540.05780.26820.15270.37120.18080.48170.2063
CrossCBR0.08260.04380.12260.05490.17860.06750.27780.16390.38150.19410.49610.2205
IDBR0.08670.04530.13410.05810.20010.07240.30820.18330.40740.21040.53010.2379
%Improv.4.9%3.4%9.3%5.8%12.0%7.2%10.9%11.8%6.8%8.3%6.8%7.9%

Ablation analysis.

MethodNetEaseYouShu
R@20N@20R@20N@20
Note: Underlined data represent the best results in the ablated models, and bold data represent the best results including IDBR.
IDBRgcn0.07910.04200.28730.1737
IDBR/ae0.07850.04070.28850.1740
IDBR/pp0.07960.04280.28930.1752
IDBR/cc0.07100.03760.27940.1676
IDBRoc0.04730.02380.24360.1358
IDBRof0.07390.03910.28780.1672
IDBR0.08670.04530.30820.1833