Optimised NN-based flood prediction on river morphological changes in the Ganga River

Floods have become a geo-environmental hazard, which has turned out to be a disaster as it has a destructive effect on the economy and society. Nowadays, several flood prediction systems have evolved that consider the previous flood events, from which the upcoming flood events can be predicted. This work develops a new flood prediction model on river morphological changes, particularly in the Ganga River. The suggested design is divided into two main parts: 1) feature extraction; 2) classification. At first, the input is given for the feature extraction phase, where the vegetation index features and water index features are extracted. After this, the extracted vegetation and water index features are classified, where an optimised deep neural network (DNN) is deployed. Furthermore, the DNN weights are adjusted using the self-improved lion algorithm (SI-LA) to increase the created approach's accuracy.