Flood risk in mountainous settlements: A new framework based on an interpretable NSGA-II-GB from a point-area duality...

Global climate change has significantly increased the frequency and intensity of extreme precipitation events, thereby heightening flood risks for mountainous settlements. However, due to geographical and socio-economic constraints in these regions, flood-related sample data are generally scarce. This study introduces a Mean Filter (MF) grounded in the point-area duality perspective, combined with a feature selection approach utilizing multi-objective optimization algorithms. Three such algorithms (NSGA-II, SPEA2, and NSPSO) were applied to generate the Pareto front of flood susceptibility features, followed by a comprehensive evaluation using the CRITIC-TOPSIS method. The results show that NSGA-II performed best in terms of Hypervolume, Spacing, Generational Distance, and the number of Pareto solutions. The selected feature subsets demonstrated improvements across multiple model validations, with the GB model achieving the highest overall performance and stability. Notably, half of the retained features are related to buffer zone and watershed characteristics, and the SHAP analysis indicates that the mean Topographic Wetness Index within the 50-m buffer zone is the main contributor to flood risk. This finding validates the effectiveness of the MF strategy based on the point-area duality perspective under limited sample conditions, offering a new approach for flood studies with sparse data in mountainous regions.