Reliable Artificial Intelligence Approach for Sustainable Flood Susceptibility Forecasting

This paper proposes a reliable approach based on machine learning techniques to enhance flood forecasting. Flood forecasting is affected by various factors including human-related factors and environmental factors. Human-related factors such as urbanization and infrastructures such as dam quality, and environmental factors such as weather data, intensity and duration of monsoon rains, and climate change effects such as deforestation. By integrating these factors, the proposed approach aims to capture the complex dynamics of flood events, enabling more accurate forecasts. The proposed approach integrates various machine learning models including Support Vector Regression (SVR), Multi-layer Perceptron (MLP), Linear Regression (LR), and Decision Tree (DT). From these models, average forecasts are calculated for the two models with the best performance. The proposed approach achieved Mean Absolute Error (MAE) of 0.0056, Mean Squared Error (MSE) of 5.431 \(\times {e}^{-5}\) , , Root Mean Squared Error (RMSE) of 0.0074, and \({R}^{2}\) of 0.9783, demonstrating its efficacy in forecasting upcoming floods effectively.