The Design and Implementation of Knowledge Graph Based Intelligent Health Question Answering System

In the context of rapid socioeconomic development, this study proposes a health intelligent question-answering system that integrates natural language processing (NLP), knowledge graphs, and large language models (LLMs) to address the limitations of traditional health management methods. The system incorporates multi-source data to construct a Neo4j-based knowledge graph, adopts a hierarchical B/S architecture for seamless multi-terminal access, and refines the Kimi model via a three-layer fine-tuning strategy (domain adaptation, label enhancement, and long-context training). Experimental results indicate that the system achieves substantial improvements in reasoning accuracy within health-related domains compared to general LLMs, with user satisfaction reaching high levels. Additionally, it enhances the efficiency of chronic disease management and improves the precision of sports injury prevention.