The Predictive Capability of a Novel Ensemble Tree-Based Algorithm for Assessing Groundwater Potential

Understanding the potential groundwater resource distribution is critical for sustainable development, conservation, and management strategies. This study analyzes maps in Busan Metropolitan City, South Korea, using random forest (RF), gradient boosting machine (GBM), extreme (XGB) methods. Fourteen conditioning factors were evaluated their contribution to assessment an elastic net. Curvature, stream power index, distance from drainage, lineament density, fault density excluded subsequent analysis, while nine other used create (GMPs) RF, GBM, XGB models. The accuracy of resultant GPMs was tested receiver operating characteristic curves seed cell area results compared. analysis showed that three models this satisfactorily predicted spatial area. In particular, model highest prediction (0.818), followed by GBM (0.802) RF (0.794). model, which most recently developed technique, found best contribute improving GPMs. These establishment a plan resources