Answers are grounded in this paper and cite their sources.
Flooding poses serious risks to agriculture, infrastructure, and human life, making accurate flood susceptibility assessment essential for effective management. This study focuses on the Cachar district of Assam in the Barak River basin, where flood events have intensified in recent years. To enhance prediction accuracy, Classification and Regression Tree (CART) models were integrated with nature-inspired optimization algorithms: Shark Smell Optimization (SSO-CART), Spotted Hyena Optimization (SHO-CART), and Wild Horse Optimization (WHO-CART). Sixteen flood-conditioning factors were initially analyzed, with twelve key variables selected for model development using feature selection techniques. Model performance was evaluated using the Area Under the Receiver Operating Characteristics (AUROC) curve. All models demonstrated strong predictive ability, with AUROC values above 0.70. Among them, WHO-CART achieved the highest accuracy, with an AUROC of 0.94. The Wilcoxon Signed-Rank test confirmed statistically significant differences between the models. Flood Susceptibility Maps indicated lower risk in the northern region and higher vulnerability in the western and central areas of Cachar. Approximately 45% of the basin (about 1,700 km²) falls within moderate to high flood-prone zones. Overall, this study presents an interpretable hybrid modeling framework that improves CART performance and supports informed flood risk management and planning.
No takes yet. Share an insight, caveat, or question.
Mishra et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: