Track: Case Studies

Abstract

Mining companies require metal concentration analysis for ore bodies such as copper, which can be costly and time-consuming, potentially negatively impacting production due to increased turnaround times. This research’s aim is to explore machine learning on geochemical data and to evaluate how machine learning methods perform in predicting metal concentrations in copper deposits. The research uses geochemistry dataset comprised of 3282 samples from the Kombat Copper deposit area in Namibia to predict copper (Cu) concentrations from zinc (Zn) and lead (Pb) concentrations. In addition to the metal concentrations, the dataset had sample coordinates and grid names. The four machine learning algorithms used were Random Forest (RF), K-Nearest Neighbor (KNN), Decision Tree (DT), and Support Vector Machine (SVM). These models were used because they were the commonly employed models for the similar purpose from literature. The machine learning model's performance was assessed using the regression score (R-squared), which quantifies the model's ability to explain data variance. Other metrics like Mean Squared Error, Root Means Squared Error, Mean Absolute Error, Adjusted R-squared, and explained variance were also considered. The KNN model outperformed the other three models, predicting 57% of the relationship between the dependent and independent variables. Further optimization of the KNN model improved the R-squared to 64% (0.64) with n-estimators set at 4. Setting the test size to 10% resulted in a 70% R-squared value (0.70). Predicting metal contents from geochemistry data with machine learning can help mining companies reduce costs by supplementing lab-based analyses with model-based predictions in determining grades.