基于文本挖掘的化工事故致因网络分析  

Network Analysis on Causes for Chemical Accidents Based on Text Mining

作　　者：李莉[1,2] 张远进[1,2] 李晓荣 LI Li;ZHANG Yuanjin;LI Xiaorong(China Research Center for Emergency Management,Wuhan University of Technology,Wuhan 430070,China;不详)

机构地区：[1]武汉理工大学中国应急管理研究中心,湖北武汉430070 [2]武汉理工大学安全科学与应急管理学院,湖北武汉430070 [3]斯旺西大学工程学院,斯旺西英国SA28PP

出　　处：《武汉理工大学学报（信息与管理工程版）》2022年第4期637-643,655,共8页Journal of Wuhan University of Technology：Information & Management Engineering

基　　金：湖北省自然科学基金青年项目(2021CFB017).

摘　　要：为明确化工生产事故的致因因素及其之间的联系,并提出针对性措施,选取2010—2020年国内200例化工事故调查报告。借助文本挖掘和社会网络分析方法,以Python和Pajek为平台,对事故报告进行分词处理,运用词频-逆文档频率(TF-IDF)算法,挖掘了事故中的38项事故特征,包括33项事故致因和5项事故类型。采用Apriori算法挖掘化工事故致因之间的强关联规则,并将事故特征绘制成网络结构图并进行社会网络分析。通过网络中心性分析和凝聚特性分析,挖掘出化工事故关键致因,以及各事故因素之间的交叉影响关系,为化工事故的预防控制提供参考。In order to clarify the causes of chemical accidents and their relationship,and propose targeted measures,200 domestic chemical accident investigation reports from 2010 to 2020 were selected.With the help of text mining and social network analysis,Python and Pajek platforms were used for word segmentation processing of accident reports.Applying TF-IDF algorithm,thirty-eight accident characteristics were excavated,including 33 accident causes and 5 accident types.The Apriori algorithm was used to obtain the strong association rules between the accident causes and network structure was mapped to proceed the social network analysis.Through the network centrality and aggregation subgroup analysis,the key chemical accident cause and the cross relationship between accident factors were dig out for reference of prevention and control of chemical accidents.

关 键 词：文本挖掘 事故致因 社会网络分析(SNA) 化工事故 关联分析 

分 类 号：X937[环境科学与工程—安全科学]