Application of bagging, boosting and stacking ensemble and EasyEnsemble methods to landslide susceptibility mapping in the Three Gorges Reservoir area of China

Abstract. Since the impoundment of the Three Gorges Reservoir area in 2003, the potential risks of geological disasters in the reservoir area have increased significantly, among which the hidden dangers of landslides are particularly prominent. To reduce casualties and damage, efficient and precise landslide susceptibility evaluation methods are important. Multiple ensemble models have been used to evaluate the susceptibility of the upper part of Badong County to landslides. In this study, EasyEnsemble technology was used to solve the imbalance between landslide and nonlandslide sample data. The extracted evaluation factors were input into three ensemble models, bagging, boosting, and stacking models, for training, and landslide susceptibility maps (LSMs) were drawn. According to the importance analysis, the important factors affecting the occurrence of landslides are altitude, terrain surface texture (TST), distance to residents, distance to rivers and land use. Comparing the influences of different grid sizes on the susceptibility results, a larger grid was found to lead to the overfitting of the prediction results. Therefore, a 30 m grid was selected as the evaluation unit. The accuracy rate, area under the curve (AUC), recall rate, test set precision, and Kappa coefficient of the multigrained cascade forest (gcForest) model under the stacking method were 0.958, 0.991, 0.965, 0.946, and 0.91, respectively, which were significantly better than the values produced by the other two models.