EFFICIENT AND RELIABLE PREDICTION OF DUMP SLOPE STABILITY IN MINES USING MACHINE LEARNING: AN IN-DEPTH FEATURE IMPORTANCE ANALYSIS.

This study rigorously examines the pressing issue of dump slope stability in Indian opencast coal mines, a problem that has led to significant safety incidents and operational hindrances. Employing machine learning algorithms such as Random Forest (RF), k-Nearest Neighbors (KNN), Support Vector Machine (SVM), Logistic Regression (LR), Decision Tree (DT), and Gaussian Naive Bayes (GNB), the study aims to achieve a scientific goal of predictive accuracy for slope stability under various environmental and operational conditions. Promising accuracies were attained, notably with RF (0.98), SVM (0.98), and DT (0.97). To address the class imbalance issue, the Synthetic Minority Oversampling Technique (SMOTE) was implemented, resulting in improved model performance. Furthermore, this study introduced a novel feature importance technique to identify critical factors affecting dump slope stability, offering new insights into the mechanisms leading to slope failures. These findings have significant implications for enhancing safety measures and operational efficiency in opencast mines, not only in India but potentially globally. [ABSTRACT FROM AUTHOR]