化工事故案例关键信息抽取研究

针对当前大部分化工事故案例因以电子文档形式存储而不利于人们对案例信息利用的问题,提出了一种化工事故案例关键信息抽取方法。该方法利用中国科学院计算所ICTCLAS(Institute of Computing Technology,Chinese Lexical Analysis System)系统对化工案例文本进行中文分词和词性标注,利用化工案例信息表述特点及句子语法信息制定关键信息识别规则,利用visual studio平台编制测试代码,对收集到的2000年至今的100个化工事故案例进行测试评价,结果显示该方法能抽取出化工事故案例发生的时间、地域、化工设备及事故类型等信息。该工作为提高化工事故案例信息的利用率作了有效的探索与尝试。

Study on Key Information Extraction of Chemical Accident Cases

A method of extracting key information of chemical accident cases is proposed because the electronic document saved currently in chemical accident cases is not conducive to the use of case information. In this method, Institute of Computing Technology, Chinese Lexical Analysis System is used to mark chemical case texts with Chinese participle and part of speech, the key information identification rules are formulated by using the characteristics of chemical case information and sentence grammar information, the visual studio platform is applied to compile test code and 100 chemical accident cases collected since 2000 have been tested and evaluated. The results show that this method can extract information such as the time, region, chemical equipment and accident type of chemical accident cases. This work makes an effective exploration and attempt to improve the utilization rate of chemical accident case information.