A comparative analysis of data mining techniques for agricultural and hydrological drought prediction in the eastern Mediterranean

•Most stations showed a significant negative (p < 0.05) trend of the standardized precipitation index (SPI) −9, −12, and −24.•The bagging (BG), random subspace (RSS), random tree (RT), and random forest (RF) algorithms were implemented for agricultural and hydrological drought prediction.•In the training stage, random tree (RT) outperformed the other algorithms (RMSE = 0.3, r = 0.97).•In the testing stage, the bagging (BG), random subspace (RSS), and random forest (RF) performance was more satisfactory than that of RT.•In the validation stage, bagging (BG) performance was satisfactory (RMSE = 0.62–0.83, r = 0.58–0.79).