Identifying and Aggregating Disaster-related Messages from Social Media Streams

Abstract

Natural and man-made disasters both cause serious casualties and economic damages Therefore one of the most challenging tasks for agencies is to rapidly and accurately identify the latest status of a disaster event With the recent widespread of social media (e g Twitter Facebook and PTT ) people may easily disseminate and share what they see and hear when they witness a severe accident These massive and instant messages containing texts photos and video may form a new type of channel that usually spreads information faster than traditional media do In this work we thus propose a scheme that improves the effectiveness of identifying disaster-related messages from social media streams Also we further attempt to aggregate messages describing an identical event This is crucial in helping disaster management agencies to understand the situation in a more comprehensive way

Date of Award	2018 Sep 1
Original language	English
Supervisor	Wei-Guang Teng (Supervisor)