Scholarlink

<p>The rapid proliferation of social media and location-based services has amplified the demand for intelligent, personalized point-of-interest (POI) recommendations. This paper introduces an advanced AI-driven framework that integrates large language models (LLMs) with deep sequential modeling to enhance sentiment-aware POI recommendation systems. By leveraging cutting-edge neural architectures, including long short-term memory (LSTM) networks and bidirectional encoder representations from transformers (BERTs), our approach effectively captures user interactions, evolving preferences, and contextual factors. A core innovation of our model lies in the application of LLM-powered sentiment analysis on user-generated reviews, extracting nuanced sentiments, implicit feedback, and fine-grained user intent. This enriched understanding enables the dynamic refinement of recommendations, ensuring relevance, diversity, and personalization. Additionally, our framework incorporates temporal dependencies, spatial context, and demographic factors, further enhancing the adaptability and accuracy of recommendations. We validate our approach through extensive experiments on two real-world location-based social network datasets, demonstrating substantial performance improvements over traditional recommendation models. Our findings indicate significant gains in precision, recall, F1-score, and user satisfaction, underscoring the impact of sentiment-aware deep learning methodologies in POI recommendation systems. This research provides valuable insights into the evolving landscape of AI-powered personalization and context-aware recommender systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An LLM-driven context-aware recommendation system integrating NLP for enhanced social media personalization

  • 期刊论文

  • 发表日期：

    2026年2月23日

查看全文
  • Qiwei Yang,
  • Wei Sun,
  • Mostafa Habibi,
  • Ibrahim Albaijan

摘要

The rapid proliferation of social media and location-based services has amplified the demand for intelligent, personalized point-of-interest (POI) recommendations. This paper introduces an advanced AI-driven framework that integrates large language models (LLMs) with deep sequential modeling to enhance sentiment-aware POI recommendation systems. By leveraging cutting-edge neural architectures, including long short-term memory (LSTM) networks and bidirectional encoder representations from transformers (BERTs), our approach effectively captures user interactions, evolving preferences, and contextual factors. A core innovation of our model lies in the application of LLM-powered sentiment analysis on user-generated reviews, extracting nuanced sentiments, implicit feedback, and fine-grained user intent. This enriched understanding enables the dynamic refinement of recommendations, ensuring relevance, diversity, and personalization. Additionally, our framework incorporates temporal dependencies, spatial context, and demographic factors, further enhancing the adaptability and accuracy of recommendations. We validate our approach through extensive experiments on two real-world location-based social network datasets, demonstrating substantial performance improvements over traditional recommendation models. Our findings indicate significant gains in precision, recall, F1-score, and user satisfaction, underscoring the impact of sentiment-aware deep learning methodologies in POI recommendation systems. This research provides valuable insights into the evolving landscape of AI-powered personalization and context-aware recommender systems.