COMPARATIVE STUDY OF EMBEDDING-BASED RECOMMENDATION METHODS FOR PERSONALIZED NEWS

Abstract

The rapid expansion of digital news platforms has significantly increased the volume of available information, creating challenges for users in identifying relevant content. In this context, recommender systems have become essential tools for filtering information and delivering personalized news feeds. However, the news domain presents unique difficulties, including high content turnover, short relevance periods, and constantly evolving user interests.

References

1. H. B. Dergi and M. B. Akgün, “Social and Categorical Signals in Contrastive Learning for Recommendation Systems,” in 2023 31st Signal Processing and Communications Applications Conf. (SIU), Istanbul, Turkiye, 2023, pp. 1–4. URL: https://doi.org/10.1109/SIU59756.2023.10223958
2. J. Yu et al., “Collaborative Filtering Recommendation with Fluctuations of User’ Preference,” in 2021 IEEE Int. Conf. on Information Communication and Software Engineering (ICICSE), Chengdu, China, 2021, pp. 222–226. URL: https://doi.org/10.1109/ICICSE52190.2021.9404120
3. J. Neve, “Content-Based Filtering,” in Reciprocal Recommender Systems, SpringerBriefs in Computer Science. Cham: Springer, 2025. URL: https://doi.org/10.1007/978-3-031-85103-2_4
4. Y. Chen et al., “Contrastive Graph Learning for Session-based Recommendation,” in 2022 4th Int. Conf. on Artificial Intelligence and Advanced Manufacturing, Hamburg, Germany, 2022, pp. 139–142. URL: https://doi.org/10.1109/AIAM57466.2022.00033
5. K. Smelyakov, A. Chupryna, D. Darahan, and S. Midina, “Effectiveness of modern text recognition solutions and tools for common data sources,” in CEUR Workshop Proc., vol. 2870, pp. 154–165, 2021. [Online]. URL: https://ceur-ws.org/Vol-2870/