事件挖掘综述  

Event Detection Overview

作　　者：李博[1] 易东云[1] 赵城利[1] 

机构地区：[1]国防科技大学数学与系统科学系,湖南长沙410073

出　　处：《计算机与现代化》2010年第11期1-4,8,共5页Computer and Modernization

基　　金：国家重点基础规划项目;即973项目(2005CB321800)

摘　　要：事件挖掘就是从大量的、不完全的、人们事先不知道的,但同时又是潜在有用的信息中提取有用信息并概括为事件的过程。由于事件挖掘相对于话题检测与跟踪、信息检索、信息抽取等自然语言处理具备许多共性,并且具备预测事件发展的优越性,因此逐渐成为当前事件识别领域的研究热点。本文介绍事件挖掘相关研究的历史发展与应用背景,全新定义事件挖掘的概念与任务,并详细介绍事件挖掘任务相关的各种关键算法,尤其是语义链算法。指出事件挖掘的重要意义及其未来的发展趋势。Event detection is the course of extracting useful information from a large amount of incomplete,unknown in advance,but potentially useful information,it is summarized into event.Since the event detection has many common attributes to topic detection and tracking,information retrieval,information extraction and other natural language processing,and has the superiority of predicting the development of event,it gradually becomes a research hotspot of current event identifying.This paper introduces the research background,task definition and related algorithms of the event detection,especially the lexical semantic chaining＇s correlation algorithm.After analyzing the research situation of current event detection field,points out the trend of development in future.

关 键 词：事件挖掘 语义链 聚类 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]