Assessment of mine slopes stability conditions using a decision tree approach

Continuous assessment of slope stability is important to the open pit design and operation. This article aims present a tool for evaluating conditions rock slopes in mining, based on global geotechnical database, using machine learning techniques. Different models are evaluated this research: general model, which uses all variables; mathematical only variables selected by random forest (out-of-bag); two expert-based models: Q-Slope model Santos model. The validation was done through test sample, partition confusion matrices aiming at reproducibility results. A study types errors carried out Principal Component Analysis (PCA). allowed identification samples that were inconsistent with others. Afterwards, redone compared previous ones. best performers presented discussed. proposed methodology does not replace classic analysis stability. On contrary, it contributes engineers geologists monitoring mining Slope must be throughout mine's lifetime and, therefore, believed here can optimize selection most susceptible instability.