Prediction of flood susceptibility in an inter-fluvial region of Northern India using machine learning algorithms

Floods are the topmost alarming hydrometeorological calamities around the globe. The Ganga-Yamuna interfluve region faces several flood hazards due to its topographical and environmental conditions. In modern times, the application of advanced technology has been implemented to predict flood susceptible regions and it predicts accurately. The principal objective of this study is to predict flood susceptible regions of the Prayagraj district of North India using advanced machine-learning models based on assessing critical flood causative factors. In addition, support vector machine (SVM), random forest (RF), extreme gradient boosting (XGBoost) and logistic regression (LR) have been applied based on fifteen topographical, hydrological, and environmental variables. The result indicates that about 15% of the area comes under high to very high flood susceptible regions. The area under the curve (AUC) result indicates that AUC values of RF, SVM, XGBoost, and LR are 0.84, 0.79, 0.85, and 0.94 respectively. The outcomes will be helpful for local administrators to take necessary action for hazard mitigation planning in flood-prone regions.