A review of text mining applications in coal mineaccident analysis

TAN Xuhong; LIU Mingming; MA He; HUANG Zhihao

doi:10.13202/j.cnki.cer.2026.01.22

Abstract：As a key technology for big data analysis,text mining has been increasingly applied in coal mine accident research in recent years,providing methodological support for accident cause identification and risk prevention.To clarify the research progress in coal mine accident analysis,this study systematically reviews and synthesizes the development and current state of text mining applications in this field,assessing various analytical approaches and models used for identifying accident causes and preventing risks.First,the concept of coal mine accident analysis is defined.The scope of the analysis is then determined by retrieving relevant literature based on document types,such as journal articles and dissertations,and combining that with their funding support status.Then,the general research process of text mining in the context of coal mine accidents is outlined,with a focus on reviewing applications across three themes:cause identification of coal mine accidents based on text features,discovery of latent knowledge in accidents using topic models,and analytical models for accidents based on multi-method integration.The review indicates that a fundamental theoretical framework has been established in this field,yet significant limitations remain in areas such as data fusion,semantic depth,and dynamic early warning.Therefore,future work should focus on developing large language models with domain adaptation,advancing technologies for collaborative analysis of multimodal data,and building intelligent early-warning systems with real-time reasoning capabilities.These developments will facilitate a transition in coal mine safety management from post-event statistics to pre-incident warning,enabling more precise and intelligent safety decision-making.