A Framework for Real-Time Disaster Warning Using Ensemble Learning, IoT sensing and Crowdsourcing

Smart cities require effective disaster management, as it directly impacts people's lives. The key challenges of disaster management are timely detection and effective notification. This research presents a smart framework for notifications’ classification and management during flooding disasters. The framework includes an early detection module as the first step in the alerting process. We propose an Ensemble Learning model (based on a triad of Deep Learning, Random Forest and K-nearest Neighbor (KNN)) to analyze data collected in real-time from Internet of Things (IoT)-based monitoring platforms. Furthermore, we propose a lightweight text-mining algorithm for crowdsourcing social data to identify the most affected areas in need of rescue. The framework integrates a fog computing layer to enable the processing of user responses in real-time and generate specialized alerts based on contextual factors like location, time, risk level, alert type, and user characteristics. Additionally, the framework proposes a centralized control panel and a smartphone application to offer essential services and facilitate communication among managed civil defense teams, citizens, and volunteers. Through testing and implementation, the proposed algorithms demonstrated an accuracy rate of over 98% in detecting threats using a real dataset of weather and flooding.