Figure 3 – Random Forest (RF) and Generalized Additive (GA) ML Models Figure 4 – Extreme Gradient Boosting (XGB) and LightGBM (LGB) ML Models Figure 5 – K-Nearest Neighbors (KNN) and Support Vector Regression (SVR) ML Models 3. RESULTS The integrated and standardized workflow successfully generates mineral prospectivity models for the Arequipa region using a normalized 500 × 500 m spatial grid. The unified feature space (0–1 scaled variables) enabled consistent comparison among heuristic and machine learning methods. The comparative analysis (Table 1) revealed clear performance differentiation among models. The Support Vector Regression (SVR) model achieved the highest explanatory capacity
RkJQdWJsaXNoZXIy MTM0Mzk2