Authors: R.V. Maitak, P.A. Pylov, E.A. Nikolaeva, A.V. Diaghileva

Title of the article: Conceptualizing the architecture of a machine learning model for classifying occupational injuries in mining enterprises ac-cording to their severity

Year: 2025, Issue: 6, Pages: 145-156

Branch of knowledge: 2.8.8 Geotechnology, mining machines

Index UDK: 622.831

DOI: 10.26730/1999-4125-2025-6-145-156

Abstract: Mining companies are production complexes that develop deposits and/or process minerals. As in any other production enterprise, work in a mining complex is associated with a number of risks and hazards that are the main cause of accidents. In order to ensure normal working conditions, each company has an industrial safety officer who ensures compliance with the safety regulations of the production process. Although the main task of these specialists is to prevent industrial accidents, they are also responsible for process safety in an emergency situation. With the advent of digital transformation, many process and industrial safety tasks have been automated (including intelligent systems) to enable more accurate preventive prediction of industrial accidents. However, in addition to preventing accidents and incidents, there is also the task of assigning an accident severity class in situations where an accident has already occurred. The lesser popularity of this topic (compared to preventive prediction of mining equipment accidents) is no reason not to automate this process. Since the severity class of an accident is assigned by a human (based on a certain set of parameters describing the accident that has occurred), this means that there is still a possibility of human error, which can only be compensated by a software model. The gaps in this area of knowledge should undoubtedly be filled by new studies and models that would allow the severity class of an accident to be determined without human intervention. The above material determines the relevance of the proposed research topic based on the automation of the process of determining the severity class of an accident in mining enterprises. The subject of the study is the process of determining the severity class of an accident using a known set of precedents describing the accident. The subject of the study is the development of a deep learning model for the analysis of a descriptive set of precedents.

Key words: mine coal industry data mining intelligent data analysis intelligent language models data processing by deep and machine learning models

Receiving date: 22.09.2025

Approval date: 15.11.2025

Publication date: 22.12.2025

CITE DOWNLOAD