A Study on Professional Technical Support for Beginners Utilizing Ontology

This study proposes a professional technical support system for beginners utilizing ontology to enhance their interaction with generative AI. The system addresses the challenge that beginners struggle to formulate appropriate prompts for AI systems. By employing ontology-based question generation, the system clarifies users’ situations through structured questioning, enabling accurate document retrieval in RAG systems. The approach extracts relevant concepts from architectural ontology and ranks questions using information entropy. Optimization techniques are applied to improve question choice, ensuring comprehensive document coverage while avoiding redundant options. Experimental results demonstrate superior performance compared to conventional RAG and long-context LLM approaches.