A New Method of Automatic Content Analysis in Disaster Management

Can A. B., PARLAK İ. B., ACARMAN T.

10th International Symposium on Digital Forensics and Security (ISDFS), Maltepe, Türkiye, 6 - 07 Haziran 2022 identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/isdfs55398.2022.9800778
  • Basıldığı Şehir: Maltepe
  • Basıldığı Ülke: Türkiye
  • Anahtar Kelimeler: topic modelling, text processing, natural language processing, smart city, deep learning, SOCIAL MEDIA, TWITTER
  • Galatasaray Üniversitesi Adresli: Evet

Özet

This study proposes a new approach to investigate the social media for disaster management. Twitter usage during an earthquake becomes a multimodal backbone in order to share the knowledge through the different aspects of the disaster. Planning the emergencies is the bottleneck of the rescue organizations in time-limited rescue intervention. Exploring the general population in the epicenter of earthquake would provide vital knowledge in rescue planning. Social media is considered as a common critical source of human information during the power outage. In this study, we focused on the analysis of rescue and non rescue topics for the 2020 Izmir earthquake. Our method analysis revealed the most important disaster topics that can be derived so that rescue organizations can successfully utilize such data. Our results provide insights into the spatio-temporal distribution of earthquake rescue/non rescue terms to identify Twitter-based discussions related to the 2020 Izmir earthquake.