摘要
为了将灾害应急任务从非结构化的文本型应急预案中自动化提取出来,提高灾害应急决策系统的自动化程度,提出了一种利用词共现模型将灾害应急任务从应急预案文本中提取出来的方法。即利用关联规则挖掘中的频繁项集抽取方法,提取应急案例文本中的共现词集,分析解释共现词集与应急任务之间的关系,并将有意义的共现词集作为提取应急任务的规则知识,应用于应急任务的提取中。以突发性大气环境污染事故应急预案为例,对该方法进行了实验,结果表明该方法能将突发性大气环境污染事故应急预案中主要的应急任务提取出来,较好地满足突发性大气环境污染应急预案应急任务的提取工作。
In order to extract the emergency task from the unstructured text type of emergency plan and improve the degree of the emergency decision system automation, a new method based on the word cooccurrence model to extract the emergency task from the type of textual emergency plan was proposed, that is, the frequent item set extraction method in association rule miming was used to extract the cooccurrence word sets from emergency case text, and to analyze and interpret the relationship between the cooccurrence word sets and emergency task. In addition, the signification of cooccurrence word sets will be selected as a rule of extracting the emergency task from text emergency plan. The method was applied to the extraction of emergency task from the emergency plan of the unexpected air environment pollution accidents. The application results show that this method can be applied to extract emergency task from the emergency plan of the unexpected air environment pollution accidents.
作者
邬群勇
方祺伟
孙振海
WU Qunyong FANG Qiwei SUN Zhenhai(Spatial Information Research Center of Fujian Province, Fuzhou University, Fuzhou 350002, China Ministry of Science and Technology, Academy of Military Medical Science, Beijing 100071, China)
出处
《自然灾害学报》
CSCD
北大核心
2016年第5期1-8,共8页
Journal of Natural Disasters
基金
国家863计划(2012AA022007)
福建省科技计划重大项目(2014H6011)~~
关键词
灾害应急任务
应急预案
多词共现
关联规则挖掘
文本挖掘
disasters emergency task
emergency plan
multiword cooccurrence
association rules mining
text mining
