Healthy Food Recommendation System Using LLM

Abstract

Healthy eating plays an important role in preventing chronic diseases and maintaining overall well-being. However, many individuals experience difficulties in selecting foods that match their nutritional needs, health conditions, and dietary goals. This study proposes a Healthy Food Recommendation System that utilizes the QWEN Large Language Model (LLM) to provide personalized food recommendations through an interactive conversational interface. The system integrates user profile information, nutritional intake tracking, and artificial intelligence-based consultation to support informed dietary decision-making. This paper applies a user-centered approach that combines nutritional monitoring and LLM-driven recommendation generation. The system records daily calorie, protein, carbohydrate, and fat consumption while allowing users to consult an AI-powered Chat Coach for personalized dietary guidance. To evaluate system reliability, we conducted negative testing on four critical functional components, including authentication validation, user profile validation, nutritional data validation, and AI query restriction validation. The experimental results demonstrate that the proposed system successfully handled all invalid scenarios and achieved a 100% pass rate across all test cases. These findings indicate that the proposed system maintains data integrity, enforces functional requirements, and provides reliable user interactions. The integration of QWEN LLM further enhances the system by delivering personalized and context-aware food recommendations. Therefore, the proposed approach offers a practical solution for intelligent dietary assistance and healthy lifestyle management.