Optimizing Rotation Forest-Based Decision Tree Algorithms for Groundwater Potential Mapping

Groundwater potential mapping is an important prerequisite for evaluating the exploitation, utilization, and recharge of groundwater. The study uses BFT (best-first decision tree classifier), CART (classification regression tree), FT (functional trees), EBF (evidential belief function) benchmark models, RF-BFTree, RF-CART, RF-FT ensemble models to map groundwater Wuqi County, China. Firstly, select sixteen spring-related variables, such as altitude, plan curvature, profile slope angle, aspect, stream power index, topographic wetness sediment transport normalized difference vegetation land use, soil, lithology, distance roads, rivers, rainfall, make a correlation analysis these variables. Secondly, optimize parameters seven optimal modeling in County. predictive performance each model was evaluated by estimating area under receiver operating characteristic (ROC) curve (AUC) statistical index (accuracy, sensitivity, specificity). results show that have good capabilities, has larger AUC value. Among them, RF-BFT highest success rate (AUC = 0.911), followed (0.898), RF-CART (0.894), (0.852), (0.824), (0.801), BFtree (0.784), respectively. maps 7 were obtained, four different classification methods (geometric interval, natural breaks, quantile, equal interval) used reclassify obtained GPM into 5 categories: very low (VLC), (LC), moderate (MC), high (HC), (VHC). breaks method best performance, most reliable. highlights proposed more efficient accurate mapping.