为深入探究煤矿瓦斯爆炸事故发展规律,提出一种基于中文分词技术对瓦斯爆炸事故情景要素进行分析和提取的方法。 通过煤矿安全网等途径搜集统计了 1978—2020 年间的 733 起瓦斯爆炸事故报告,在此基础上进行数据预处理,剔除不完整的事故报告,最终选取 255 起瓦斯爆炸事故报告进行要素分析与提取;将事故等级、事故经过、事故原因等内容进行整理储存,形成待挖掘文本语料库;基于 Jieba 分词算法提取瓦斯爆炸事故情景关键词,并采用 TF-IDF 算法进行权重计算,将情景划分为事故体、致灾体、承灾体、抗灾体 4 个维度和 24 个要素,为后续瓦斯爆炸事故的情景表示和事故未来的可能性组合提供了参考依据。

In order to deeply explore the development law of coal mine gas explosion accidents,a method to analyze and extract the scenario elements of gas explosion accidents based on Chinese word segmentation technology was proposed. First of all,733 gas explosion accident reports from to were collected and counted through the coal safety network and other means.On this basis,data preprocessing was carried out. Incomplete accident reports were excluded,and gas explosion accident reports were finally selected for the element analysis and extraction. The classification,process and cause of accidents were sorted and stored to form a corpus of texts to be mined. Keywords of gas explosion accident scenario were extracted based on Jieba word segmentation algorithm,and TF - IDF algorithm for weight calculation. The scenarios were divided into dimensions and elements,including accident body,disaster - causing body,disaster- bearing body and disaster - resistant body, which provided a reference basis for the scenario representation of the subsequent gas explosion accidents and the possible cause combination for the future possibilities of the accidents.