Study on classification of coal mine accident causes based on NLP

doi:10.16265/j.cnki.issn1003-3033.2023.06.1416

Abstract

Abstract:

In order to improve the efficiency of analyzing and processing coal mine accident text effectively, NLP and accident cause models were used for building an automatic classification framework of accident cause. Based on 24Model, 87 typical mine accident investigation reports were analyzed, and a framework of mine accident cause classifications was obtained. A corpus was constructed for each type of accident cause constructed. The NLP was used for processing each type of cause text of the corpus and training the fastText model to realize the automatic recognition and classification of accident cause text. The method proposed was compared with TextCNN and the other two classical models. The results show that a total of 21 types of accident causes and 6 684 training corpus are obtained, the accuracy of fastText after training can reach 98.92%, and the comprehensive performance is better than the other three methods. The accident text mining system is developed based on 24Model and NLP, which can analyze and process the accident text information quickly and further detail the cause of the accident investigation report, which is convenient for the case study and statistical analysis.

Key words: natural language processing(NLP), classification of accidents causes, "2-4″model(24Model), fastText

ZHANG Jiangshi, LI Yongtun, MAO Xiangning, HU Xinyue, WANG Ziyi. Study on classification of coal mine accident causes based on NLP[J]. China Safety Science Journal, 2023, 33(6): 20-26.

Figures/Tables 9

Fig.1

Fig.2

Tab.1

Fig.3

Fig.4

Fig.5

Tab.2

Fig.6

Fig.7

References 21

[1]	陈晓. 基于数据挖掘的煤矿安全管理知识可视化研究[D]. 北京: 中国矿业大学(北京), 2017.
	CHEN Xiao. Research on visualization of coal mine safety management knowledge based on data mining[D]. Beijing: China University of Mining and Technology-Beijing, 2017.
[2]	CHOI Y G, CHO K T. Analysis of safety management characteristics using network analysis of CEO messages in the construction industry[J]. Sustainability, 2020: DOI:10.3390/su12145771. doi: 10.3390/su12145771
[3]	郭对明, 李国清, 胡乃联, 等. 基于文本挖掘的矿山安全隐患大数据分析与可视化[J]. 工程科学学报, 2022, 44(3): 328-338.
	GUO Duiming, LI Guoqing, HU Nailian, et al. Big data analysis and visualization of mine safety hazards based on text mining[J]. Journal of Engineering Science, 2022, 44(3):328-338.
[4]	LU Dan, XU Changqing, MI Chuanmin, et al. Establishment of a key hidden danger factor system for electric power personal casualty accidents based on text mining[J]. Information, 2021: DOI:10.3390/info12060243. doi: 10.3390/info12060243.
[5]	NAKATA T. Extracting typical incident patterns from text data[J]. International Journal of Human Factors Modelling and Simulation, 2018, 6(2/3): 127-139. doi: 10.1504/IJHFMS.2018.093177
[6]	武佳豪. 基于文本的企业安全隐患关联规则挖掘及预警方法[D]. 北京: 中国石油大学(北京), 2020.
	WU Jiahao. Text-based mining and warning method of enterprise security risks association rules[D]. Beijing: China University of Petroleum-Beijing, 2020.
[7]	李擎, 张秋艳, 白磊. 一种基于文本挖掘的铁路基础设施设备风险隐患识别模型[J]. 铁路计算机应用, 2018, 27(2):1-4.
	LI Qing, ZHANG Qiuyan, BAI Lei. An identification model of railway infrastructure equipment risk and hidden danger based on text mining[J]. Railway Computer Application, 2018, 27(2):1-4.
[8]	韩天园, 田顺, 吕凯光, 等. 基于文本挖掘的重特大交通事故成因网络分析[J]. 中国安全科学学报, 2021, 31(9):150-156. doi: 10.16265/j.cnki.issn1003-3033.2021.09.021
	HAN Tianyuan, TIAN Shun, LYU Kaiguang, et al. Cause network analysis of major traffic accidents based on text mining[J]. China Safety Science Journal, 2021, 31(9):150-156. doi: 10.16265/j.cnki.issn1003-3033.2021.09.021
[9]	陈曼玲. 基于文本挖掘的建筑工人不安全行为研究[D]. 武汉: 华中科技大学, 2020.
	CHEN Manling. Research on construction workers' unsafe behavior based on text mining[D]. Wuhan: Huazhong University of Science and Technology, 2020.
[10]	GOLDBERG David M. Characterizing accident narratives with word embeddings: improving accuracy, richness, and generalizability[J]. Journal of Safety Research, 2022, 80: 441-455. doi: 10.1016/j.jsr.2021.12.024 pmid: 35249625
[11]	MORAIS C, YUNG K L, JOHNSON K, et al. Identification of human errors and influencing factors: a machine learning approach[J]. Safety Science, 2022:DOI:10.1016/j.ssci.2021.105528. doi: 10.1016/j.ssci.2021.105528.
[12]	ZHANG Fan, FLEYEH H, WANG Xinru, et al. Construction site accident analysis using text mining and natural language processing techniques[J]. Automation in Construction, 2019, 99: 238-248. doi: 10.1016/j.autcon.2018.12.016
[13]	煤矿安全网. 煤矿事故/事故案例[EB/OL]. [2022-07-22]. https://www.mkaq.org/sggl/shigual/.
[14]	JOULIN A, GRAVE E, BOJANOWSKI P, et al. Bag of tricks for efficient text classification[C]. Association for Computational Linguistics, 2017: 427-431.
[15]	MIKOLOV T, CHEN Kai, CORRADO G, et al. Efficient estimation of word representations in vector space[J]. Computer Science, 2013:DOI:10.48550/arXiv.1301.3781. doi: 10.48550/arXiv.1301.3781
[16]	傅贵, 王秀明, 李亚. 事故致因“2-4”模型及其事故原因因素编码研究[J]. 安全与环境学报, 2017, 17(3): 1003-1008.
	FU Gui, WANG Xiuming, LI Ya. Research on the "2-4" model of accident causation and its coding of accident causation factors[J]. Journal of Safety and Environment, 2017, 17(3): 1003-1008.
[17]	搜狗输入法. 词库/工程与应用科学[EB/OL]. [2022-07-11]. https://pinyin.sogou.com/dict/.
[18]	闫琰, 杨梦, 周法国, 等. 基于fastText网络的煤矿事故案例文本分类方法对比[J]. 煤炭工程, 2021, 53(11):186-192.
	YAN Yan, YANG Meng, ZHOU Faguo, et al. Comparison of text classification methods of coal mine accident cases based on fastText network[J]. Coal Engineering, 2021, 53(11): 186-192.
[19]	GOH Y M, UBEYNARAYANA C U. Construction accident narrative classification: an evaluation of text mining techniques[J]. Accident Analysis & Prevention, 2017, 108: 122-130. doi: 10.1016/j.aap.2017.08.026
[20]	煤矿事故/事故案例/《陕西省神木市百吉矿业有限责任公司“1·12”重大煤尘爆炸事故调查报告》[EB/OL]. [2022-07-11].https://www.mkaq.org/html/2021/09/17/583306.shtml.
[21]	吕千, 傅贵. 事故调查报告中事故原因分析的改进:基于企业事故案例学习视角[J]. 中国安全生产科学技术, 2021, 17(10): 172-178.
	LYU Qian, FU Gui. Improvement of accident cause analysis in accident investigation report: from the perspective of enterprise accident case study[J]. Journal of Safety Science and Technology, 2021, 17(10): 172-178.

24Model事故原因分类体系			细化后的分类
组织内因素	直接原因	不安全动作	1.违章指挥
			2.违规作业
			3.操作失误
		不安全物态	4.不安全物态
	间接原因	安全知识不足	5.安全知识不足
		安全意识不高	6.安全意识不高
		安全习惯不佳	7.安全习惯不佳
		心理状态不佳	8.心理状态不佳
		生理状态不佳	9.生理状态不佳
	根本原因	安全管理体系欠缺	10.未建立体系文件
			11.体系文件不健全
			12.设计缺陷
			13.配套设施不齐全
			14.人员组织不合理
			15.制度落实不到位
			16.安全培训不到位
			17.隐患排查不到位
	根源原因	安全文化欠缺	18.安全文化欠缺
组织外因素			19.政府部门疏于监管
			20.第三方问题
			21.其他因素

模型	A /%	P /%	R /%	F₁ /%	运行时间/s
fastText	98.92	98.58	99.09	98.84	1.9
SVM	98.82	99.43	93.26	94.60	8.28
NBC	98.49	99.43	93.75	94.78	0.17
TextCNN	98.49	98.82	96.63	97.49	176.83