基于文本挖掘的高校实验室事故致因识别研究

Research on Identifying the Causes of Accidents in University Laboratories Based on Text Mining

摘要: 为有效预防和减少高校实验室事故的发生率，使用ROST CM6文本挖掘工具和UCINET社会网络分析工具，基于从各网络、媒体收集的2006~2022年间的高校实验室事故报道26份，从词频分析、共现矩阵分析、事故致因可视化处理、中心性分析四个角度切入，探索不同事故致因的出现频率和事故致因之间的相互关系。研究结果显示：违规操作致因出现的频率最大，为35.71%；安全管理体系缺失致因、安全意识不足致因与违规操作致因共现次数最多，达3次；违规操作致因属于事故致因网络的中心位置且中心度最高，与其他致因关联性最强，是最中心的事故致因。

Abstract: For the purpose of guarding effectively against and decreasing the incidence of accidents in university laboratories, using the ROST CM6 content mining system and UCINET social network recognition tool, based on 26 university laboratory accident reports collected from various networks and media from 2006 to 2022, this study explores the risk patterns and interrelationships from four perspectives: word frequency analysis, co-occurrence matrix analysis, accident cause visualization processing, and centrality analysis. The results showed that the frequency of causes of violations was the highest, at 35.71%. The most common causes of the lack of safety management system, insufficient safety awareness, and violation of regulations were three times. The cause of illegal operations belongs to the central position of the accident cause network, with the highest centrality and the strongest correlation with other causes, making it the most central accident cause.