Tsunamis are among the natural disasters that cause maximum destruction and devastation of both life and property; an early prediction of such events can help mitigate their impacts. The existing methods for tsunami prediction suffer some serious limitations in accuracy and real-time adaptability. In this line, we propose a hybrid tsunami prediction framework based on signal processing, IoT data, and advanced time-series modelling using a SARIMAX-GRU model. The workflow involves data collection, pre-processing, feature extraction, and model training, to ensure effective prediction of tsunami events. It’s astonishing accuracy and same precision, recall, and F1 score of 98.25% place the proposed model among the best in predicting tsunami occurrences, when compared to other models such as SVM, RF, and KNN. The contribution of this work will offer tsunami prediction and risk management researchers a comprehensive approach in integrating the seismic and hydrological data with advanced deep learning methodologies.
Muniasamy et al. (Fri,) studied this question.