Earthquake Damage Assessment Based on User Generated Data in Social Networks

摘要: Natural disasters have always been one of the threats to human societies. As a result such crises, many people will be affected, injured, and financial losses incur. Large earthquakes often occur suddenly; consequently, crisis management is difficult. Quick identification affected areas after critical events can help relief workers provide emergency services more quickly. This paper uses social media text messages create damage map. A support vector machine (SVM) machine-learning method was used identify mentions among messages. The map created based on damage-related tweets. results showed SVM classifier accurately identified where F-score attained 58%, precision 56.8%, recall 59.25%, accuracy 71.03%. In addition, temporal pattern non-damage tweets investigated each day per hour. analysis that most were sent earthquake. our research evaluated by comparing with official intensity maps. findings earthquake estimated efficiently strategy at multispatial units an overall 69.89 spatial grid unit Spearman’s rho Pearson correlation 0.429 0.503, respectively, county unit. We two in this examine impact assessment. determine priority workers.