The Cold Start Problem Solution Using BERT model in Recommendation Systems
DOI:
https://doi.org/10.59743/jbs.v38i1.327Keywords:
BERT Model, Cold Start Problem, Neural Collaborative Filtering (NCF)Abstract
Recommendation systems are crucial technologies in many modern applications such as e-commerce, social networks, and digital streaming platforms like Netflix and YouTube. These systems provide personalized suggestions to users based on their past interaction data or the characteristics of the content they engage with. One of the most significant challenges facing these systems is the Cold Start Problem, which occurs when it is difficult to provide accurate recommendations due to limited or new user interaction data or item information. This issue becomes more complex when dealing with new users or unknown items,To address this problem, the BERT deep learning model was utilized to enhance the system's ability to provide accurate recommendations based on textual features of users, in addition to leveraging demographic data such as age and gender. This approach has proven effective in improving the performance of these systems, achieving a text classification accuracy of up to 96%, demonstrating the model's strong generalization capabilities on new data. Furthermore, the BERT model showed superior efficiency in this study compared to the Neural Collaborative Filtering (NCF) model.
References
Y. Shi, M. Larson, and A. Hanjalic, “Collaborative filtering beyond the user-item matrix: A survey of the state of the art and future challenges,” ACM Comput. Surv., vol. 47, no. 1, pp. 1–45, Jan. 2014.
Y. Koren, R. Bell, and C. Volinsky, “Matrix factorization techniques for recommender systems,” Computer, vol. 42, no. 8, pp. 30–37, Aug. 2009.
Y. Luo, Y. Jiang, Y. Jiang, G. Chen, J. Wang, K. Bian, ... and Q. Zhang, "Online Item Cold-Start Recommendation with Popularity-Aware Meta-Learning," arXiv preprint arXiv:2411.11225, 2024.
B. He, J. Zhou, R. Zhao, L. Wang, L. Li, and Y. Yang, "Cold-Start Product Recommendation Method Based on GAE," Journal of Computing and Electronic Information Management, vol. 13, no. 3, pp. 8–15, 2024.
X. Wu, H. Zhou, Y. Shi, W. Yao, X. Huang, and N. Liu, "Could Small Language Models Serve as Recommenders? Towards Data-centric Cold-start Recommendations," arXiv preprint arXiv:2306.17256, 2024.
X. Lin, W. Wang, J. Zhao, Y. Li, F. Feng, and T.-S. Chua, "Temporally and Distributionally Robust Optimization for Cold-Start Recommendation," in Proceedings of the 38th AAAI Conference on Artificial Intelligence (AAAI-24), 2024.
Y. Li, Y. Liu, and T. Furukawa, "Integrating Prior Knowledge from Meta-Learning and Large Language Models for Cold-Start Recommendation," in Proceedings of International Exchange and Innovation Conference on Engineering & Sciences (IEICES), 2023, pp. 327-333.
L. Zheng, J. Chen, P. Liu, G. Zhang, and J. Fang, "VM-Rec: A Variational Mapping Approach for Cold-start User Recommendation," arXiv preprint arXiv:2307.04709, 2023.
B. Hao, H. Yin, J. Zhang, C. Li, and H. Chen, "A Multi-Strategy based Pre-Training Method for Cold-Start Recommendation," Proc. ACM Meas. Anal. Comput. Syst., vol. 37, no. 4, pp. 1–24, May 2022.
M. Volkovs, G. Yu, and T. Poutanen, "DropoutNet: Addressing Cold Start in Recommender Systems," Proceedings of the 31st International Conference on Neural Information Processing Systems (NeurIPS), Long Beach, CA, USA, 2017, pp. 4003–4013.
C.W. Leung, S.C. Chan and K.K.F. Chung. Applying Cross-Level Association Rule Mining to Cold-Start Recommendations, Proceedings of the 2007 IEEE/WIC/ACM International Conference on Web Intelligence and International Conference on Intelligent Agent Technology, 2007, Silicon Valley, CA, USA, pag: 133-136
A. Vaswani, “Attention is all you need,” in Advances in Neural Information Processing Systems, 2017.
J. D. M. W. C. Kenton and L. K. Toutanova, “BERT: Pre-training of deep bidirectional transformers for language understanding,” in Proceedings of NAACL-HLT, vol. 1, no. 2, Jun. 2019.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Journal of Basic Sciences
This work is licensed under a Creative Commons Attribution 4.0 International License.
