Comprehensive Comparative Analysis of the Personalization Diversity Trade-off in News Recommendation Systems
Deep Learning Model Evaluation on MIND Dataset with Hybrid Framework Optimization
DOI:
https://doi.org/10.63635/mrj.v2i2.285Keywords:
News Recommendation Systems, Personalization-Diversity Trade-off, Deep Learning, MIND Dataset, Filter Bubbles, Hybrid Recommender Systems, Multi-Objective OptimizationAbstract
News Recommendation Systems (NRS) face a fundamental challenge: optimizing for personalization accuracy inevitably exacerbates filter bubbles by reducing content diversity. This study presents a comparative evaluation of five models on the Microsoft News Dataset (MIND) containing 1 million users and 24.2 million interaction logs. We evaluate Content-Based Filtering (CBF), Collaborative Filtering (CF), Deep Neural Networks (DNN), Recurrent Neural Networks (RNN) and Attention-Based Neural Networks (ABNN) using twelve metrics spanning personalization accuracy, diversity and system-level properties. Results demonstrate that ABNN achieves superior personalization (Precision@10 = 0.401) but severely reduces diversity (ILD = 0.387), a 47.6% decrease from CBF's diversity baseline (ILD = 0.738). We propose a novel hybrid framework incorporating diversity-aware loss functions and multi-objective optimization that achieves 91% of maximum personalization while improving diversity by 45.2% at less than 5% computational overhead. Our findings indicate that practitioners should deliberately occupy specific Pareto-optimal positions based on use-case requirements and that filter bubble effects disproportionately harm specialist users while intensifying over time.
Downloads
References
[1] EY; Media and entertainment report. Invest India 2023, https://static.investindia.gov.in/s3fs-public/2023-05/ey-me-report.pdf
[2] Liu, J.; Dolan, P.; Pedersen, E.R.; Personalized news recommendation based on click behavior. Proc. 15th Int. Conf. Intell. User Interfaces (IUI), Hong Kong, China 2010, 31–40, https://doi.org/10.1145/1719970.1719976 DOI: https://doi.org/10.1145/1719970.1719976
[3] Pariser, E.; The Filter Bubble: What the Internet is Hiding from You. Penguin Press: New York, 2011. DOI: https://doi.org/10.3139/9783446431164
[4] Sunstein, C.R.; #Republic: Divided Democracy in the Age of Social Media. Princeton University Press: Princeton, 2017. DOI: https://doi.org/10.1515/9781400884711
[5] Resnick, P.; Iacovou, N.; Suchak, M.; Bergstrom, P.; Riedl, J.; GroupLens: An open architecture for collaborative filtering of netnews. Proc. ACM Conf. Comput. Supported Coop. Work (CSCW), Chapel Hill, NC, USA 1994, 175–186, https://doi.org/10.1145/192844.192905 DOI: https://doi.org/10.1145/192844.192905
[6] Koren, Y.; Bell, R.; Volinsky, C.; Matrix factorization techniques for recommender systems. Computer 2009, 42(8), 30–37, https://doi.org/10.1109/MC.2009.263 DOI: https://doi.org/10.1109/MC.2009.263
[7] Covington, P.; Adams, J.; Sargin, E.; Deep neural networks for YouTube recommendations. Proc. 10th ACM Conf. Recomm. Syst. (RecSys) 2016, 191–198, https://doi.org/10.1145/2959100.2959190 DOI: https://doi.org/10.1145/2959100.2959190
[8] Liu, Z.; Xing, Y.; Wu, F.; An, M.; Xie, X.; Hi-Fi Ark: Deep user representation via high-fidelity archive network. Proc. 28th Int. Joint Conf. Artif. Intell. (IJCAI) 2019, 3059–3065, https://doi.org/10.24963/ijcai.2019/424 DOI: https://doi.org/10.24963/ijcai.2019/424
[9] Hidasi, B.; Karatzoglou, A.; Baltrunas, L.; Tikk, D.; Session-based recommendations with recurrent neural networks. Proc. Int. Conf. Learn. Represent. (ICLR) 2016, https://doi.org/10.48550/arXiv.1511.06939
[10] Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A.N.; Kaiser, Ł.; Polosukhin, I.; Attention is all you need. Adv. Neural Inf. Process. Syst. (NeurIPS) 2017, 30, 5998–6008, https://doi.org/10.48550/arXiv.1706.03762
[11] Ziegler, C.-N.; McNee, S.M.; Konstan, J.A.; Lausen, G.; Improving recommendation lists through topic diversification. Proc. 14th Int. Conf. World Wide Web (WWW) 2005, 22–32, https://doi.org/10.1145/1060745.1060754 DOI: https://doi.org/10.1145/1060745.1060754
[12] Ge, M.; Delgado-Battenfeld, C.; Jannach, D.; Beyond accuracy: Evaluating recommender systems by coverage and serendipity. Proc. 4th ACM Conf. Recomm. Syst. (RecSys) 2010, 257–260, https://doi.org/10.1145/1864708.1864761 DOI: https://doi.org/10.1145/1864708.1864761
[13] Zheng, Y.; Gao, C.; Chen, L.; Jin, D.; Li, Y.; DGCN: Diversified recommendation with graph convolutional networks. Proc. Web Conf. (WWW) 2021, 401–412, https://doi.org/10.1145/3442381.3449835 DOI: https://doi.org/10.1145/3442381.3449835
[14] Wu, F.; Qiao, Y.; Chen, J.-H.; Wu, C.; Qi, T.; Lian, J.; Liu, D.; Xie, X.; Gao, J.; Wu, W.; Zhou, M.; MIND: A large-scale dataset for news recommendation. Proc. 58th Annu. Meet. Assoc. Comput. Linguist. (ACL) 2020, 3597–3606, https://doi.org/10.18653/v1/2020.acl-main.331 DOI: https://doi.org/10.18653/v1/2020.acl-main.331
[15] Pazzani, M.J.; Billsus, D.; Content-based recommendation systems. In The Adaptive Web, Lecture Notes in Computer Science, vol. 4321; Springer: Berlin, Heidelberg, 2007, 325–341, https://doi.org/10.1007/978-3-540-72079-9_10 DOI: https://doi.org/10.1007/978-3-540-72079-9_10
[16] Wu, C.; Wu, F.; Ge, S.; Qi, T.; Huang, Y.; Xie, X.; Neural news recommendation with multi-head self-attention. Proc. Conf. Empir. Methods Nat. Lang. Process. Int. Joint Conf. Nat. Lang. Process. (EMNLP-IJCNLP) 2019, 6389–6394, https://doi.org/10.18653/v1/D19-1671 DOI: https://doi.org/10.18653/v1/D19-1671
[17] Steck, H.; Calibrated recommendations. Proc. 12th ACM Conf. Recomm. Syst. (RecSys) 2018, 154–162, https://doi.org/10.1145/3240323.3240372 DOI: https://doi.org/10.1145/3240323.3240372
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Dhanjit Kalita, Tribikram Pradhan, Dhritiman Saikia (Author)
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Copyright © Author(s) retain the copyright of this article.
