Embedded Feature Selection and Machine Learning Methods for Flash Flood Susceptibility-Mapping in the Mainstream Songhua River Basin, China

Mapping flash flood susceptibility is effective for mitigating the negative impacts of floods. However, a variety conditioning factors have been used to generate maps in various studies. In this study, we proposed combining logistic regression (LR) and random forest (RF) models with embedded feature selection (EFS) filter specific sets two map mainstream basin Songhua River. According EFS results, optimized included 32 28 features LR RF models, respectively, composition optimal was similar distinct. Overall, relevant vegetation cover river exhibit relatively high effects overall floods study area. The provided accurate reliable (FFSMs). model (accuracy = 0.8834, area under curve (AUC) 0.9486) better prediction capacity than 0.8634, AUC 0.9277). Flash flood-prone areas are mainly distributed south southwest close rivers. results obtained useful prevention control projects.