Knowledge Graph-Enhanced Chinese Intelligent Agent for Tumor Rehabilitation

This study develops a knowledge graph-based system for tumor exercise rehabilitation, integrating retrieval-augmented generation (RAG) to enhance question-answering capabilities. The system leverages large language models (LLMs), knowledge graphs to establish a complete reasoning process, from data acquisition to precise rehabilitation recommendations. It adopts a three-tier architecture comprising the perception layer, logic layer, and application layer, enabling efficient data processing and prescription reasoning. The frontend is implemented using Python, while the backend employs MySQL for relational data storage and Neo4j for graph-based data management.In the tumor exercise rehabilitation knowledge graph visualization module, this study details the construction of the display interface, retrieval module, and storage update mechanism. The retrieval-augmented generation (RAG) question-answering module includes a retrieval component, a scoring mechanism, and a historical record storage module, facilitating accurate and context-aware responses.