Authors : Navin Kumar Sehgal; Antim Dev Mishra

Volume/Issue : Volume 10 - 2025, Issue 12 - December

Google Scholar : https://tinyurl.com/3x4ra6jh

Scribd : https://tinyurl.com/3uks8u6b

DOI : https://doi.org/10.38124/ijisrt/25dec1061

Note : A published paper may take 4-5 working days from the publication date to appear in PlumX Metrics, Semantic Scholar, and ResearchGate.

Note : Google Scholar may take 30 to 40 days to display the article.

Abstract : Recommender systems traditionally rely on historical interaction data and latent-factor models, which often fail to capture users’ dynamic intentions, contextual preferences, and short-term goals. This study proposes a hybrid recommendation framework that integrates Large Language Model (LLM)-derived user control variables with matrix factorization to improve prediction accuracy and model responsiveness. Using natural-language prompts, the LLM extracts four structured control features—Perspective, Variation, Organizing, and Restore—which represent user intent, exploration preference, active interest clusters, and noise reduction signals. These control variables are fused with user and item latent factors through a control-aware rating function that adjusts the baseline matrix factorization output. Experimental evaluation on the Book-Crossing dataset demonstrates that incorporating LLM-derived controls reduces RMSE by up to 7.8% and increases Precision by 12.3% compared to standard matrix factorization. Additional analysis shows improved robustness against noisy historical data and increased alignment between recommended items and users stated short-term objectives. The findings highlight the effectiveness of semantic user-control extraction in enhancing recommender accuracy and provide a scalable path for integrating intent-aware mechanisms in modern personalized

Keywords : Recommender Systems, Large Language Models (LLMs), User Control Signals, Matrix Factorization, Hybrid Recommendation Framework, Intent-Aware Personalization, Preference Modeling, Semantic Feature Extraction, Accuracy Enhancement, User-Centric Recommendation, Context-Aware Recommendations, Control-Driven Recommenders.

References :

1. Y. Koren, R. Bell, and C. Volinsky, “Matrix factorization techniques for recommender systems,” Computer, vol. 42, no. 8, pp. 30–37, 2009.
2. G. Adomavicius and A. Tuzhilin, “Context-aware recommender systems,” AI Magazine, vol. 32, no. 3, pp. 67–80, 2011.
3. R. Burke, “Hybrid recommender systems: Survey and experiments,” User Modeling and User-Adapted Interaction, vol. 12, no. 4, pp. 331–370, 2002.
4. T. B. Brown, B. Mann, N. Ryder, et al., “Language models are few-shot learners,” Advances in Neural Information Processing Systems, vol. 33, pp. 1877–1901, 2020.
5. J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “BERT: Pre-training of deep bidirectional transformers for language understanding,” arXiv preprint arXiv:1810.04805, 2018.
6. C.-N. Ziegler, “Book-Crossing Dataset,” University of Freiburg, Germany, 2004.
7. Y. Hu, Y. Koren, and C. Volinsky, “Collaborative filtering for implicit feedback datasets,” Proc. IEEE Int. Conf. Data Mining, pp. 263–272, 2008.
8. J. L. Herlocker, J. A. Konstan, L. G. Terveen, and J. T. Riedl, “Evaluating collaborative filtering recommender systems,” ACM Transactions on Information Systems, vol. 22, no. 1, pp. 5–53, 2004.
9. D. Jannach, A. Manzoor, M. Oulabi, and M. Ashok, “A survey on conversational recommender systems,” ACM Computing Surveys, vol. 54, no. 5, pp. 1–38, 2021.
10. OpenAI, “GPT-4 technical report,” arXiv preprint arXiv:2303.08774, 2023.
11. S. Zhang, L. Yao, A. Sun, and Y. Tay, “Deep learning based recommender system: A survey and new perspectives,” ACM Computing Surveys, vol. 52, no. 1, pp. 1–38, 2019.
12. F. M. Harper and J. A. Konstan, “The MovieLens datasets: History and context,” ACM Transactions on Interactive Intelligent Systems, vol. 5, no. 4, Article 19, 2016.