A Multi-Session Framework for Adaptive Personalization in Context-Aware and Real-Time Recommender Systems
Keywords:
Next-Item Prediction, Transformer Self-Attention, Context-Aware Personalization, Graph Neural Networks (GNNs), Session-Based RecommendationAbstract
This study presents a hybrid session-based recommendation framework that integrates a Context-Aware Gated Graph Neural Network (CA-GGNN) with a Transformer encoder to enhance personalization in dynamic, anonymous user environments. The model addresses key limitations of traditional session-based recommender systems, which often disregard contextual factors, struggle with long-range dependencies, and fail to generalize in sparse or complex sessions. In this study, each user session is represented as a graph structure capturing item transitions, enriched with contextual inputs such as time, device, and interaction intervals. The CA-GGNN component performs gated message passing to learn structural dependencies while integrating session-level context, whereas the Transformer encoder models global sequential behaviour through multi-head self-attention. A fusion layer combines graph-based and sequential representations to generate a unified embedding used for next-item prediction. Experiments on the Yoochoose dataset demonstrate that CA-GGNN substantially outperforms SR-GGNN, and that the hybrid CA-GGNN + Transformer model achieves further gains, particularly in long and complex sessions. Evaluation using P@20 and MRR@20 confirms improved ranking quality, while qualitative case studies highlight the hybrid model’s ability to interpret shifting user intent more effectively. Overall, the proposed framework provides a scalable, context-sensitive, and real-time solution for next-item prediction in modern recommendation environments.
References
A. B. da Silva and E. J. Spinosa, “A Sequential Recommender System with Embeddings Based on GraphSage Aggregators,” Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 13654 LNAI, pp. 1–15, 2022, doi: 10.1007/978-3-031-21689-3_1.
A. Maatouk, F. Ayed, N. Piovesan, A. De Domenico, M. Debbah, and Z. Q. Luo, “TeleQnA: A Benchmark Dataset to Assess Large Language Models Telecommunications Knowledge,” IEEE Netw., 2025, doi: 10.1109/MNET.2025.3576035.
B. Wu and Y. Zhu, “Denoising dual sparse graph attention model for session-based recommendation,” Discov. Comput., vol. 28, no. 1, pp. 1–17, Dec. 2025, doi: 10.1007/S10791-025-09532-2/FIGURES/4.
C. Kumar, M. Abuzar, and M. Kumar, “MGU-GNN: Minimal Gated Unit based Graph Neural Network for Session-based Recommendation,” Appl. Intell., vol. 53, no. 20, pp. 23147–23165, Oct. 2023, doi: 10.1007/S10489-023-04679-1/METRICS.
C. Liu, Y. Li, H. Lin, and C. Zhang, “GNNRec: gated graph neural network for session-based social recommendation model,” J. Intell. Inf. Syst., vol. 60, no. 1, pp. 137–156, Feb. 2023, doi: 10.1007/S10844-022-00733-5/METRICS.
D. Jin et al., “Dual Intent Enhanced Graph Neural Network for Session-based New Item Recommendation,” ACM Web Conf. 2023 - Proc. World Wide Web Conf. WWW 2023, pp. 684–693, Apr. 2023, doi: 10.1145/3543507.3583526.
F. Wang, X. Gao, Z. Chen and L. Lyu, "Contrastive Multi-Level Graph Neural Networks for Session-Based Recommendation," in IEEE Transactions on Multimedia, vol. 25, pp. 9278-9289, 2023, doi: 10.1109/TMM.2023.3250087.
G. Sun et al., “OL-JCMSR: A Joint Coding Monitoring Strategy Recommendation Model Based on Operation Log,” Math. 2022, Vol. 10, Page 2292, vol. 10, no. 13, p. 2292, Jun. 2022, doi: 10.3390/MATH10132292.
H. Li, C. Çelik, M. Bashir, L. Zou, and Z. Yang, “Incorporation of a global perspective into data-driven analysis of maritime collision accident risk,” Reliab. Eng. Syst. Saf., vol. 249, p. 110187, Sep. 2024, doi: 10.1016/J.RESS.2024.110187.
J. Dąbrowski et al., “An Efficient Manifold Density Estimator for All Recommendation Systems,” Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 13111 LNCS, pp. 323–337, 2021, doi: 10.1007/978-3-030-92273-3_27.
J. Kim, J. Kim, and J. Choi, “Sequential Movie Genre Prediction Using Average Transition Probability with Clustering,” Appl. Sci. 2021, Vol. 11, Page 11841, vol. 11, no. 24, p. 11841, Dec. 2021, doi: 10.3390/APP112411841.
J. Luo, M. He, W. Pan, and Z. Ming, “BGNN: Behavior-aware graph neural network for heterogeneous session-based recommendation,” Front. Comput. Sci., vol. 17, no. 5, pp. 1–16, Oct. 2023, doi: 10.1007/S11704-022-2100-Y/METRICS.
J. Wang, H. Xie, F. L. Wang, L. K. Lee, and M. Wei, “Jointly modeling intra- and inter-session dependencies with graph neural networks for session-based recommendations,” Inf. Process. Manag., vol. 60, no. 2, p. 103209, Mar. 2023, doi: 10.1016/J.IPM.2022.103209.
J. Wang, L. K. Lee, and N. I. Wu, “Dual-Channel Convolutional Recurrent Networks for Session-Based Recommendation,” Lect. Notes Networks Syst., vol. 370, pp. 287–296, 2022, doi: 10.1007/978-981-16-8664-1_25.
J. Yu, C. Pan, Y. Li, and J. Wang, “An Academic Text Recommendation Method Based on Graph Neural Network,” Inf. 2021, Vol. 12, Page 172, vol. 12, no. 4, p. 172, Apr. 2021, doi: 10.3390/INFO12040172.
J. Zhang, C. Ma, X. Mu, P. Zhao, C. Zhong, and A. Ruhan, “Recurrent convolutional neural network for session-based recommendation,” Neurocomputing, vol. 437, pp. 157–167, May 2021, doi: 10.1016/J.NEUCOM.2021.01.041.
K. Narang, Y. Song, A. Schwing, and H. Sundaram, “FuseRec: fusing user and item homophily modeling with temporal recommender systems,” Data Min. Knowl. Discov., vol. 35, no. 3, pp. 837–862, May 2021, doi: 10.1007/S10618-021-00738-8/METRICS.
L. Yin, J. Lu, G. Zheng, H. Chen, and W. Deng, “Recommendation Algorithm for Multi-Task Learning with Directed Graph Convolutional Networks,” Appl. Sci. 2022, Vol. 12, Page 8956, vol. 12, no. 18, p. 8956, Sep. 2022, doi: 10.3390/APP12188956.
M. Hamada, R. Jawad, K. Almahmood, and A. Tekerek, “Issues and Solutions in Deep Learning-Enabled Recommendation Systems within the E-Commerce Field,” Appl. Sci. 2022, Vol. 12, Page 11256, vol. 12, no. 21, p. 11256, Nov. 2022, doi: 10.3390/APP122111256.
M. Marcuzzo, A. Zangari, A. Albarelli and A. Gasparetto, "Recommendation Systems: An Insight Into Current Development and Future Research Challenges," in IEEE Access, vol. 10, pp. 86578-86623, 2022, doi: 10.1109/ACCESS.2022.3194536.
M. Vikram, N. S. Reddy, and K. Madhavi, “SessionRNNRec: a deep learning based framework for modelling user sessions to generate accurate recommendations,” Int. J. Syst. Assur. Eng. Manag., pp. 1–10, Jul. 2021, doi: 10.1007/S13198-021-01197-6/METRICS.
P. Li, M. De Rijke, H. Xue, S. Ao, Y. Song, and F. D. Salim, “Large Language Models for Next Point-of-Interest Recommendation,” SIGIR 2024 - Proc. 47th Int. ACM SIGIR Conf. Res. Dev. Inf. Retr., pp. 1463–1472, Jul. 2024, doi: 10.1145/3626772.3657840.
Q. Chen, Z. Guo, J. Li, and G. Li, “Knowledge-enhanced Multi-View Graph Neural Networks for Session-based Recommendation,” SIGIR 2023 - Proc. 46th Int. ACM SIGIR Conf. Res. Dev. Inf. Retr., pp. 352–361, Jul. 2023, doi: 10.1145/3539618.3591706/SUPPL_FILE/SIGIR23-FRP1027.MP4.
Q. Shen, S. Zhu, Y. Zhang, Z. Wei, E. Khan, and M. Gönen, “Temporal aware Multi-Interest Graph Neural Network for Session-based Recommendation,” Proceedings of Machine Learning Research, vol. 189. PMLR, pp. 2022–2022, Apr. 13, 2023. Accessed: Aug. 06, 2025. https://proceedings.mlr.press/v189/shen23a.html
Rafiq, Wajid & Senhaji Hafid, Abdelhakim & Qadir, Junaid. (2023). Developing Smart City Services Using Intent-Aware Recommendation Systems: A Survey. Transactions on Emerging Telecommunications Technologies. 34. 10.1002/ett.4728.
S. Sang, N. Liu, W. Li, Z. Zhang, Q. Qin, and W. Yuan, “High-order attentive graph neural network for session-based recommendation,” Appl. Intell., vol. 52, no. 14, pp. 16975–16989, Nov. 2022, doi: 10.1007/S10489-022-03170-7/METRICS.
S. Sang, W. Yuan, W. Li, Z. Yang, Z. Zhang, and N. Liu, “Position-aware graph neural network for session-based recommendation,” Knowledge-Based Syst., vol. 262, p. 110201, Feb. 2023, doi: 10.1016/J.KNOSYS.2022.110201.
T. Huang, H. Yu, and G. Fan, “Dual-Channel Graph Contextual Self-Attention Network for Session-Based Recommendation,” Lect. Notes Inst. Comput. Sci. Soc. Telecommun. Eng. LNICST, vol. 406 LNICST, pp. 244–258, 2021, doi: 10.1007/978-3-030-92635-9_15.
T. R. Gwadabe and Y. Liu, “IC-GAR: item co-occurrence graph augmented session-based recommendation,” Neural Comput. Appl., vol. 34, no. 10, pp. 7581–7596, May 2022, doi: 10.1007/S00521-021-06859-X/METRICS.
X. Fang, “Making recommendations using transfer learning,” Neural Comput. Appl., vol. 33, no. 15, pp. 9663–9676, Aug. 2021, doi: 10.1007/S00521-021-05730-3/METRICS.
X. Gu, H. Zhao, and L. Jian, “Sequence neural network for recommendation with multi-feature fusion,” Expert Syst. Appl., vol. 210, p. 118459, Dec. 2022, doi: 10.1016/J.ESWA.2022.118459.
X. Zhang, Z. Jia, C. Ma, X. Jiang, and X. Xing, “Context-Aware Intent-Driven Session-based Model for Session-based Recommendation,” Proc. 2025 IEEE 14th Data Driven Control Learn. Syst. Conf. DDCLS 2025, pp. 837–842, 2025, doi: 10.1109/DDCLS66240.2025.11065988.
Y. Deldjoo, “Understanding Biases in ChatGPT-based Recommender Systems: Provider Fairness, Temporal Stability, and Recency,” ACM Trans. Recomm. Syst., Aug. 2024, doi: 10.1145/3690655;JOURNAL:JOURNAL:TORS;TAXONOMY:TAXONOMY:ACM-PUBTYPE;PAGEGROUP:STRING:PUBLICATION.
Y. Deldjoo, T. Di Noia, and F. A. Merra, “A Survey on Adversarial Recommender Systems: From Attack/Defense Strategies to Generative Adversarial Networks,” ACM Comput. Surv., vol. 54, no. 2, Mar. 2022, doi: 10.1145/3439729;PAGE:STRING:ARTICLE/CHAPTER.
Y. Liang, Q. Song, Z. Zhao, H. Zhou, and M. Gong, “BA-GNN: Behavior-aware graph neural network for session-based recommendation,” Front. Comput. Sci., vol. 17, no. 6, pp. 1–10, Dec. 2023, doi: 10.1007/S11704-022-2324-X/METRICS.
Y. Ni, N. Wang, L. Chen, R. Chen, and L. Li, “TUR: Utilizing Temporal Information to Make Unexpected E-Commerce Recommendations,” Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 13724 LNCS, pp. 553–563, 2022, doi: 10.1007/978-3-031-20891-1_39.
Y. Zhang, “Cross-Modal Neural Architecture Integrating BERT, 3D-CNN and DNN for Personalized Recommendation of Ideological and Political Educational Resources,” 3rd Int. Conf. Data Sci. Inf. Syst. ICDSIS 2025, 2025, doi: 10.1109/ICDSIS65355.2025.11070338.
Downloads
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
