- Conference Article
- 10.1109/icbats66542.2025.11258215
Context-Aware Product Recommendations for Emerging Markets: a Sentiment-Driven LLM Approach
- May 01, 2025
- Amrita Ahuja + 1 more +1
In today's world, personalized recommendations are crucial for maintaining and enhancing user engagements. The way users interact with when shopping online, more personalized products they see, the better the engagement rates would be for businesses. Traditionally, conventional recommendation techniques cannot effectively with finegrained semantic meaning of the user-generated content. This makes harder for businesses to engage with customers and make less precise suggestions. This research paper, focuses and presents a hybrid recommendation approach that is proposed to exploit embeddings from Large Language Models and sentiment analysis of user reviews to enhance the level of personalization. This framework is tested, fine-tuned and enhanced on simulated dataset containing 27,000 reviews and achieves noticeable enhancement in recommendation accuracy and relevance. This framework presents a scalable personalization-based recommendations solution by resolving data sparsity and noisy user feedback issues. This framework is especially useful for the emerging markets where data is limited and incorporating sentiment analysis enhances business analytics capability, resulting in richer consumer preferences and behaviors insights. This study brings forth the possibility of integrating semantic comprehension and sentiment analysis to enhance context-aware product recommendation systems, especially in emerging economies.
Read more