基于古籍疫病医案数据挖掘的AI智能体建设研究

打开文本图片集

分类号 R2-52;G254.929.1

DOI10.16810/j.cnki.1672-514X.2025.07.008

Research on the Construction of an AI Agent Based on Data Mining of Ancient Medical Records of Epidemics

Xu Jie，Liang Guoqing， Ni Liqiang， Zhang Huaiqiong，Wu Tingting

AbstractThis paper takes the ancient medical records of epidemics inthe library colection as theobject of data mining and proposes an AI paradigm based on a multi-level RAG-Rerank architecture.This paradigm innovates the traditional means of data pre-processing and methods of data mining.Furthermore，by using the LLMs application development framework （SemanticKernel），combined with software engineering methodology，and adopting advanced development technologies （Python，C#）andthe integrateddevelopment environment （Visual Studio），across-platform AI agentapplication is formed.Tests show that theAI agent effctively alleviates the cost problem caused by term standardization inthe traditional data pre-processing stage.Italsoaddresses issuessuch as model hallucination during the AI-empowerment process，abnormal accuracyof the traditional RAG low-code platform，andthe inability to customize thecore logic.In thecontext ofAI+，this paper explores afeasible cross-research method fordata mining and auxiliary diagnosis and treatment research based on ancient Chinese medical literature.

KeywordsAncient medical records.AI agent.RAG.LLMs.Data mining.Software engineerint

0引言

中华文明绵延5000余年，历久弥新；而中医药则是这漫长历史留给我们的珍贵礼物，不断传承至今。（剩余12547字）