Natural Disaster Prediction with Cloud Integration for Enhancing Validation Using Gradient Boosting Algorithm

Abstract

Annually, natural catastrophes inflict heavy tolls in terms of human lives lost, property destroyed, and economies hit hard. Information and communications technology (ICT) enables the executions of diverse models that handle different parts of disaster management, while geospatial scientists work to reduce or control these risks via computer modelling of these complicated occurrences. Complex natural hazard models, requiring large-scale computing resources, intense data, and concurrent access, make it impossible to execute these models using standard ICT foundations in a timely way. Offering natural hazard modelling systems as end service is now more feasible than ever before, because cloud computing's almost infinite computing, storage, and networking power, which can handle these issues. Researchers still have a way to go before they can fully embrace and put Cloud Computing technology to use in disaster response. Therefore, this paper compiles all these difficulties, discusses recent trends in the field, and lays out a theoretical framework for a Cloud-based solution that makes use of the Gradient Boosting algorithm to improve the validation of systems for modelling and managing natural hazards. These systems would make use of Cloud infrastructures in tandem with other technologies, such as IoT network, fog, and edge computing. Result shows that 90% of accuracy and less prediction time compared to training time of the dataset.