Reducing Uncertainty in Mineral Potential Mapping (MPM) using Ensemble Machine Learning and Average Voting

This study proposes a simple and uncertainty-aware ensemble framework for mapping Pb-Zn mineralization. Three regression-based machine learning models including Linear Regression, Decision Tree Regression, and Support Vector Regression are employed to capture different learning behaviors and geological relationships. To reduce model-related uncertainty, an Average Voting strategy is applied by averaging normalized outputs of the individual models. Prediction uncertainty is quantified based on model disagreement, allowing the identification of low-risk zones with high confidence. Results indicate that ensemble modeling improves spatial stability and reduces uncertainty. Low-risk targets derived from high average prospectivity and minimal model divergence show strong agreement with known Pb-Zn occurrences. The proposed approach provides a transparent, computationally efficient, and transferable solution for uncertainty reduction in mineral prospectivity mapping.

کلیدواژه ها:

Linear Regression ، Decision Tree Regression ، Support Vector Regression ، Mineral Prospectivity Mapping