A Language-Driven Framework for Improving Personalized Recommendations: Merging LLMs with Traditional Algorithms

2025-07-09

AI Summary

Traditional recommender systems struggle to model users’ fine-grained preferences expressed in natural language (e.g., “light-hearted and humorous comedies”). To address this, we propose a language-enhanced personalized recommendation framework: first, initial recommendations are generated via SVD or SVD++; then, a large language model (LLM) parses user-provided textual descriptions or historical interaction sequences to automatically construct a semantic interest profile. Crucially, we introduce a novel “quasi-friend” recommendation mechanism that leverages linguistic signals to refine preference understanding and optimize ranking. Experiments on the MovieLens-Latest-Small dataset demonstrate substantial improvements over baseline methods: cumulative hit rate increases by 6× and NDCG by 3.7×, confirming significant gains in both ranking quality and personalization fidelity. This work empirically validates the critical role of natural language feedback in enhancing recommendation accuracy and interpretability.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Traditional recommendation algorithms are not designed to provide personalized recommendations based on user preferences provided through text, e.g., "I enjoy light-hearted comedies with a lot of humor". Large Language Models (LLMs) have emerged as one of the most promising tools for natural language processing in recent years. This research proposes a novel framework that mimics how a close friend would recommend items based on their knowledge of an individual's tastes. We leverage LLMs to enhance movie recommendation systems by refining traditional algorithm outputs and integrating them with language-based user preference inputs. We employ Singular Value Decomposition (SVD) or SVD++ algorithms to generate initial movie recommendations, implemented using the Surprise Python library and trained on the MovieLens-Latest-Small dataset. We compare the performance of the base algorithms with our LLM-enhanced versions using leave-one-out validation hit rates and cumulative hit rates. Additionally, to compare the performance of our framework against the current state-of-the-art recommendation systems, we use rating and ranking metrics with an item-based stratified 0.75 train, 0.25 test split. Our framework can generate preference profiles automatically based on users' favorite movies or allow manual preference specification for more personalized results. Using an automated approach, our framework overwhelmingly surpassed SVD and SVD++ on every evaluation metric used (e.g., improvements of up to ~6x in cumulative hit rate, ~3.7x in NDCG, etc.), albeit at the cost of a slight increase in computational overhead.

Problem

• Enhancing personalized recommendations using LLMs and text-based preferences
• Integrating traditional algorithms with LLMs for better movie suggestions
• Automating preference profiles to improve recommendation accuracy and personalization

Innovation

• LLMs enhance traditional recommendation algorithms
• SVD/SVD++ generate initial movie recommendations
• Automated preference profiles improve personalization