基于大语言模型的信用监管领域知识图谱构建方法

Construction method of knowledge graph in credit supervision domain based on large language model

随着数字信息时代的到来,信息技术及营商环境得到了持续地发展与优化。在信用监管领域,数据具有海量、多源、异构、实体间关系复杂、隐含关系难以发现等问题。对此,文章提出了一种基于大型语言模型的领域知识图谱构建方法。实验结果表明,该方法的F1值达到89.3%,相较传统方法有显著提升。信用监管领域知识图谱的构建实现了市场主体失信信息的可视化呈现与查询,有助于揭示市场主体间的隐含关系并识别失信风险,可有效提升监管效能,助力智慧监管全面落地。

With the advent of the digital information age,information technology and business environment have been continuously developed and optimized.In the field of credit supervision,data has problems such as massive,multi-source,heterogeneous,complex relationships between entities,and difficulty in discovering implicit relationships.This article proposes a domain knowledge graph construction method based on large-scale language models.The experimental results show that the F1 value of this method reaches 89.3%,which is significantly improved compared to traditional methods.The construction of a knowledge graph in the field of credit supervision has enabled the visualization and retrieval of information on market entitiesdishonesty,which helps to reveal implicit relationships between market entities and identify dishonesty risks.This can effectively enhance regulatory efficiency and assist in the comprehensive implementation of smart supervision.