摘要
以科技文献数据库(web of science,WOS)为数据源,2015-2019年国内、外情报学领域期刊论文为研究对象,综合采用文献计量学和文献述评理论与方法,利用CiteSpace软件进行关键词共现分析,获取文本信息抽取方法的研究热点及发展路径,并对四类主要的文本信息抽取方法进行评述。通过对比分析发现以上四种方法并不是独立存在的,文本信息抽取方法呈现多元化和综合化趋势,深度学习逐渐成为最新的发展方向,知识地图成为最新的信息抽取框架。
In order to obtain research hotspots and development paths,this paper selects relevant journal papers of information science field in WOS database in 2015-2019,and uses the theory and methods of bibliometrics and literature review to conduct citation and keyword co-occurrence analysis by the CiteSpace software.Meanwhile,four main text information extraction methods are reviewed.Through comparative analysis,it is found that the above four methods do not exist independently,and the text information extraction methods are diversified and integrated.Deep learning has gradually become the latest development direction,and knowledge maps have become the latest extraction framework.
作者
蔡皎洁
Cai Jiaojie(Hubei Small and Micro Enterprise Development Research Center,Hubei Engineering University,Xiaogan,Hubei 432000,China)
出处
《湖北工程学院学报》
2019年第6期65-72,共8页
Journal of Hubei Engineering University
基金
湖北省软科学项目(2019ADC148)
湖北省社科基金项目(19ZD057)
关键词
文本信息
信息抽取
深度学习
知识地图
web information
information extraction
deep learning
knowledge graph
