Conceptual Representation for Crisis-Related Tweet Classification

Won-Gyu Choi, Kyung-Soon Lee


The importance of social media such as Twitter, as a conduit for actionable and tactical information during disasters is increasingly recognized. During crisis situations, rapid and effective response actions by emergency services are critical to assure the safety of the public. In this paper, we propose a conceptual representation for crisis-related tweet classification. In order to classify a stream of tweets related to the incident, the crisis-related terms in each tweet are represented as conceptual entities such as event entities, category indicator entities, information type entities, URL entities, and user entities. For tweet classification, we have compared support vector machines and deep learning model which combines class activation mapping with one-shot learning in convolutional neural networks. Experimental results on TREC 2018 Incident Streams test collection show significant improvement over the baseline system.


Conceptual representation, crisis-related tweets classification, incident streams, convolutional neural networks, class activation mapping, one-shot learning, support vector machines

Full Text: PDF