Role of Social Media in Crisis Response and Disaster Management During Natural Calamities

This paper proposes a disaster management system based on machine learning, incorporating a web-based graphical user interface. The system involves six steps, starting with the inclusion of any disaster-related information into the message dataset. Two commonly used machine learning algorithms, random forest and logistic regression, are employed to evaluate the performance of the interface, achieving an accuracy of 72% and 73%, respectively, in classification and prediction tasks. The interface enables users to input messages containing relevant keywords classified based on predefined criteria. Afterwards, the system automatically selects the appropriate department for the particular disaster. The proposed system can enhance disaster management by simplifying the identification and response to disasters.