Development of Disaster Management Frameworkusing Natural Language Processing

Abstract

Social networks are increasingly used for emergency communications and help related requests. During disaster situations, such emergency requests need to be mined from the pool of big data for providing timely help. The sentiment of the affected people during and after the disaster determines the success of the disaster response and recovery process. Our goal is to explore, understand the underlying trends in sentiment with respect to disasters and geographically related sentiment, helping the affected area recover soon. The proposed model DRM Framework (Disaster Recovery and Management Framework) collects disaster data from social networks and other outlets, categorizes them according to the needs of the affected people & classifies the severity level of a particular Geographic Location using our own Natural Language Processing (NLP) Model. The categorized disaster data are classified through a machine learning algorithm for analysing the sentiment of the people. The practical implication of the proposed methodology is the real-time categorization and classification of Big Data for disaster response and recovery. This analysis helps the emergency responders and rescue personnel to develop better strategies for effective information and resource management in a rapidly changing disaster environment.