基于自然语言处理的“双碳”政策知识图谱构建及应用  

Research on the construction of a knowledge graph for dual carbon policies based on natural language processing

作　　者：吕涛[1] 王青山 张紫玉 吴昱磊 周孜柔 王洛 LYU Tao;WANG Qingshan;ZHANG Ziyu;WU Yulei;ZHOU Zirou;WANG Luo(School of Economics and Management,China University of Mining and Technology,Xuzhou 221116,China)

机构地区：[1]中国矿业大学经济管理学院,江苏徐州221116

出　　处：《煤炭经济研究》2025年第2期122-132,共11页Coal Economic Research

基　　金：教育部产学合作协同育人项目(221005278104521);中国矿业大学创新创业训练项目(202410290213Y,202310290078Z)。

摘　　要：“双碳”政策具有发布数量多、覆盖范围广、内容复杂多样等特点,现有的呈现方式难以满足知识检索和内在分析的需求。以2953条“双碳”政策文本为数据源,提出了一种基于自然语言处理的“双碳”政策知识图谱构建方法,首先构建了知识图谱模式层,定义了“双碳”政策实体、属性和关系,之后采用Text Rank关键词抽取、LDA主题建模等算法提取政策实体、属性及关系,构建了知识图谱数据层,最终将〈实体,关系,实体〉三元组存入Neo4j图数据库,形成“双碳”政策知识图谱。所构建的知识图谱包含2048个实体节点和32336条关系,可通过Cypher语言实现不同细粒度政策实体和关系的关联查询与可视化,挖掘“双碳”政策中的关键语义信息和政策热点,还可为智能服务提供语义增强功能,提高“双碳”政策推荐系统的效率和政策问答系统的准确度。The"dual carbon"policy is characterized by a high frequency of releases,broad scope,and complex,multifaceted content.The existing presentation methods are difficult to meet the needs of knowledge retrieval and internal analysis.This paper takes 2953 dual carbon policy texts as the data source and proposes a dual carbon policy knowledge graph construction method based on natural language processing.First,the knowledge graph model layer is constructed,and the dual carbon policy entities,attributes and relationships are defined.Then,the Text Rank keyword extraction,LDA topic modeling and other algorithms are used to extract policy entities,attributes and relationships to construct the knowledge graph data layer.Finally,the"entity,relationship,entity"triples are stored in the Neo4j graph database to form a dual carbon policy knowledge graph.The constructed knowledge graph contains 2048 entity nodes and 32336 relationships.The Cypher language enables the implementation of association queries and visualization of various fine-grained policy entities and their relationships.Additionally,it facilitates the extraction of key semantic information and the identification of policy hotspots within the dual carbon policy framework.It can also provide semantic enhancement functions for intelligent services,improve the efficiency of the dual carbon policy recommendation system and the accuracy of the policy question-answering system.

关 键 词：“双碳”政策 知识图谱 自然语言处理 Neo4j LDA Text Rank 

分 类 号：F416.21[经济管理—产业经济]