文本数据事件检测的研究热点及趋势分析

Research Hotspots and Trend Analysis of Text Data Event Detection

随着科学技术的发展,人们发布信息、表达观点的渠道越来越多,且最常见的信息载体就是文本。文本数据事件检测能够从海量数据中识别和检测当前正在发生的热点或突发事件,可以很好地支持应急管理、舆情监控、信息安全等领域的工作。为了掌握文本数据事件检测的研究状况,揭示该领域的研究热点和发展趋势,该文以中国知网作为文献数据库,2003年-2021年的440篇期刊论文作为样本,借助科学计量软件CiteSpace,从发文时间分布、基金支持、主要研究力量、主要刊载平台、高被引文献分布、研究热点以及演化路径等方面,对文本数据事件检测研究进行计量分析。研究表明,事件检测研究发文量趋于稳定,且研究质量正在不断提高。研究热点为突发事件与热点话题的文本事件检测应用研究、基于微博数据的事件检测研究和以聚类为主要方法的事件检测方法研究。研究发展分为概念形成和工具开发两个阶段,并已经出现了领域扩散的现象。现阶段的研究重点集中在事件检测技术,社交媒体事件检测和事件检测在突发事件中的应用等方面。基于深度学习的文本数据检测方法、在社交媒体中的突发事件检测逐渐成为本领域研究的发展趋势。

With the development of science and technology,people have more and more channels to publish information and express their opinions.The most common information carrier is text.Text data event detection identifies and detects hot or unexpected events that are currently occurring from massive amounts of data,and it can well support emergency management,public opinion monitoring,information security and other fields.In order to grasp the research status of text data event detection and reveal the research hotspots and development trends in this field,we use CNKI as a literature database,440 journal articles from 2003-2021 as a sample,econometric analysis of text data event detection research in terms of distribution of publication time,funding support,major research power,major publication platforms,distribution of highly cited literature,research hotspots and evolutionary paths.The results show that the number of event detection research publications is stabilizing and the quality of research is improving.The research hotspots are the application research of text event detection for breaking news and hot topics,research on event detection based on twitter data and research on event detection method based on clustering.Research development is divided into two phases:concept formation and tool development,and domain proliferation has already occurred.At this stage,the research focus is divided into three categories,event detection technology,event detection research in social media and event detection research in emergent events.Text data detection methods based on deep learning and emergency detection in social media have gradually become the development trend of this field.