基于实体知识网络的政策文本关联分析——以国家创业政策为例
Policy Text Relevance Analysis Based on Entity Knowledge Network——Taking China's Entrepreneurship Policies as an Example
黄佳妮 1陆伟 2于丰畅3
作者信息
- 1. 武汉大学信息管理学院
- 2. 武汉大学信息管理学院;武汉大学信息检索与知识挖掘研究所
- 3. 武汉大学信息检索与挖掘研究所
- 折叠
摘要
国家政治或经济目的的达成往往需要众多政策协同实施,探究政策间的关联、厘清其关系有助于政策的科学制定和有效实践.针对现有研究对政策文本关联的语义层面揭示较浅等问题,为进一步挖掘政策之间的关联关系、发现政策群中的核心政策,文章提出一种多层次政策实体知识网络的关联挖掘框架.运用命名实体识别技术抽取政策文本中的细粒度政策知识对象,在此基础上构建多层次政策实体知识网络,设计基于Jaccard距离的关联关系定量计算方法.文章以创业政策群为例构建多层次政策实体知识网络,有效揭示政策文本间的潜在关联,发现政策群中的核心政策文本,对政策文本关系进行多维度阐释.
Abstract
The achievement of national political or economic goals often requires the coordinated implementation of all kinds of policies.Therefore,exploring the relationship and relevance among different policies is conducive to the scientific formulation and effective practice of such policies.However,current researches have done little in revealing the semantic level of policy text relevance.In order to know more about the relationship and relevance among different policies and discover the core policy in each policy group,this paper proposes a multi-level relevance mining framework of policy entity knowledge network.First,named entity recognition technologies are used to extract fine-grained policy knowledge objects in policy texts.Then,a multi-level policy entity knowledge network is constructed,and a quantitative calculation method based on Jaccard distance is designed.Afterwards,an analysis is conducted of the self-constructed a multi-level entrepreneurial policy entity knowledge network.The result shows that such a network can effectively reveal the potential relevance among different policy texts,discover the core policy in each policy group,and interpret the relationship among the policy texts from multiple dimensions.
关键词
政策关联分析/政策文本计算/知识网络/知识抽取Key words
policy relevance analysis/policy text calculation/knowledge network/knowledge extraction引用本文复制引用
基金项目
湖北省博士后创新研究岗位项目(2021)(211000090)
出版年
2024
NETL
NSTL
万方数据