Robustness of Optimized Decision Tree-Based Machine Learning Models to Map Gully Erosion Vulnerability

Gully erosion is a worldwide threat with numerous environmental, social, and economic impacts. The purpose of this research to evaluate the performance robustness six machine learning ensemble models based on decision tree principle: Random Forest (RF), C5.0, XGBoost, treebag, Gradient Boosting Machines (GBMs) Adaboost, in order map predict gully erosion-prone areas semi-arid mountain context. first step was prepare inventory data, which consisted 217 points. This database then randomly subdivided into five percentages Train/Test (50/50, 60/40, 70/30, 80/20, 90/10) assess stability models. Furthermore, 17 geo-environmental variables were used as potential controlling factors, several metrics examined results revealed that all performed well terms predicting vulnerability erosion. C5.0 RF had best prediction (AUC = 90.8 AUC 90.1, respectively). However, according random subdivisions database, these exhibit small but noticeable instability, high for 80/20% 70/30% subdivisions. demonstrates significance refining need test various splitting data ensure efficient reliable output results.