Ensemble Machine Learning for Urban Flood Hazard Assessment

Urban flood hazard assessment using an ensemble approach can reduce the bias of individual models and provide a more accurate picture of how flood risks may change in specific locations over time. By incorporating different models, the ensemble approach can produce a more accurate prediction of flooding events. In the current research, we used an ensemble machine learning for flood hazard assessment. The results showed that the ensemble model outperforms other methods, such as the classification and regression tree (CART) method and random forest (RF). The results of the hazard mapping verify the credibility of the obtained data for raising awareness and informing the public about flood-prone areas.