数字图书馆论坛2025，Vol.21Issue(10)：1-11,11.DOI:10.3772/j.issn.1673-2286.2025.10.001
融合图尔敏模型与大语言模型的社会科学循证知识库构建方法
Evidence-Based Knowledge Base Construction Method for Social Sciences Integrating Toulmin Model and LLM
摘要
Abstract
Evidence-based research in the social sciences faces dual challenges:the rapid growth of literature and complex argumentation scenarios,necessitating intelligent and efficient evidence processing methods.This study proposes a method for constructing a social sciences evidence-based knowledge base that integrates the Toulmin argumentation model with large language model(LLM),aiming to improve the efficiency of processing social sciences evidence and the level of intelligence in evidence analysis.First,we design an information extraction process centered on argumentation.Second,we construct a multidimensional knowledge organization model for semantic retrieval.Finally,we design an intelligent evidence-based service process based on chain reasoning,and take library and information science as an example to develop the prototype system ToulminQA.Experimental evaluations demonstrate that ToulminQA exhibits high accuracy and reliability in argumentative element extraction.The system effectively integrates extensive evidence to generate logically coherent,practically valuable evidence-based conclusions.This research provides a novel reference and practical example for evidence-based research pathways in the social sciences.关键词
图尔敏模型/大语言模型/社会科学/知识库构建/循证社会科学/证据/知识组织Key words
Toulmin Model/Large Language Model/Social Sciences/Knowledge Base Construction/Evidence-Based Social Sciences/Evidence/Knowledge Organization分类
社会科学引用本文复制引用
齐涵悦,杨颜僖,林泽斐..融合图尔敏模型与大语言模型的社会科学循证知识库构建方法[J].数字图书馆论坛,2025,21(10):1-11,11.基金项目
本研究得到福建省社会科学基金项目"基于大型语言模型的社会科学循证知识库构建"(编号:FJ2024BF034)资助. （编号:FJ2024BF034）
