Author ORCID Identifier

https://orcid.org/0009-0005-4886-7994

Date Available

8-12-2025

Year of Publication

2025

Document Type

Master's Thesis

Degree Name

Master of Science in Mining Engineering (MSMIE)

College

Engineering

Department/School/Program

Mining Engineering

Faculty

Pedram Roghanchi

Abstract

This thesis examines the predictive capability of a temporal machine learning model for forecasting future accidents and violations at individual mines, based on historical data. Mine accidents were categorized by accident classification and violations were categorized by the Part Section. The primary datasets utilized were the mine safety and health administration’s (MSHA’s) Accident Injuries and Violations datasets. The available datasets were cleaned and organized by mine type and commodity, then divided into separate subsets for training, validating, and testing. Different models, cutoff metrics, learning rates, number of hidden layers, data processing methods, data processing divisions, number of points observed per iteration, percentage of training dataset used per iteration, and activation function were compared using the underground coal dataset to determine which algorithms had the highest values for F1-score, precision, and recall. The highest performing algorithms based on F1-score for the accident and violation datasets were tested against the most cited accidents and violation for each mine type and commodity. Especially on high occurring accidents and violations, a temporal machine learning model can be used to assist in the prediction of such events. The majority of points did not reach satisfactory standards which led to the inclusion of the highest performing models based on precision and recall in the final predictions to generate multiple risk levels per incident. The best performing dataset was the underground coal dataset leading to the assumptions that the best performing algorithms vary depending on the dataset due to this data being used to choose the algorithms to begin with. All functions were implemented into a user-friendly interface along with the ability to display pre-existing information in the datasets. This research provides a basis for evaluating the safety performance of mining operations across the United States.

Digital Object Identifier (DOI)

https://doi.org/10.13023/etd.2025.317

Funding Information

This study was funded by the National Institute for Occupational Safety and Health (NIOSH) under the award #U60OH012685 in 2023

Download

Share

COinS