Research on the roof hazard factors quantification based on Chinese word segmentation and TF-IDF algorithm

Area of Science:

• Mining Engineering
• Occupational Safety and Health
• Data Science

Background:

• Roof accidents pose significant risks in coal mine safety.
• Existing hazard identification methods are subjective and inefficient.
• A growing research trend highlights the need for better analysis of roof accident literature.

Purpose of the Study:

• To develop an automated framework for extracting and quantifying roof hazard factors from unstructured accident reports.
• To overcome the limitations of traditional subjective and manual hazard identification techniques.
• To objectively identify and weight key direct and indirect causes of roof accidents in coal mines.

Main Methods:

• An integrated framework combining Chinese word segmentation and the TF-IDF algorithm was employed.
• The method was applied to analyze 115 unstructured roof accident reports.
• Automated extraction and quantification of hazard factors and key phrases were performed.

Main Results:

• A significant reduction in redundant information for direct (26%) and indirect (35%) causes was achieved.
• Key direct causes identified include mine worker actions, regulatory violations, and inadequate support.
• Key indirect causes identified involve insufficient safety training, lack of regulation, and poor hazard management.

Conclusions:

• The proposed framework offers an objective and efficient approach to analyzing coal mine roof hazards.
• Automated extraction and quantification of hazard factors improve the understanding of accident causation.
• This method enhances coal mine safety by providing data-driven insights into accident prevention.