Integration of statistical models and ensemble machine learning algorithms (MLAs) for developing the novel hybrid groundwater potentiality models: a case study of semi-arid watershed in Saudi Arabia

The present study has proposed three novel hybrid models by integrating traditional ensemble models, such as random forest, logitboost, and naive bayes, six newly developed of rotation forest (RF), decision tree (RF-DT), J48 (DF-J48), bayes (RF-NBT), neural network (RF-NN), M5P (RF-M5P) REPTree (RF-REPTree), with statistical i.e. weight evidence, logistic regression combination WOE LR. To predict the groundwater potential, nine potential conditioning parameters have been created. Information Gain Ratio used to evaluate impact each parameter. ROC curve validate models. According findings, 15 30% area a very high or potentiality. Furthermore, validation results revealed that RF based ensembles outperformed other standalone for modelling.