Home // HUSO 2016, The Second International Conference on Human and Social Analytics // View article

Analysing Emotions in Social Media Coverage on Paris Terror Attacks: a Pilot Study

Authors:
Cynthia Van Hee
Celine Verleye
Els Lefever

Keywords: Emotion detection; Social media; Natural language processing; Terrorism

Abstract:
Social media provide an increasingly used platform for crisis communication. Governments need to understand how publics consume and react to crisis information via social media. One option to do this is by applying emotion analysis. In this pilot study, we target the November 2015 terrorist attacks in Paris as a case study for emotion analysis and detection. We constructed a Dutch Facebook corpus manually annotated with i) Ekman’s basic emotions and ii) irony use. The annotations reveal that anger is the most recurrent emotion, however the basic emotions do not cover all emotions in the dataset. The corpus also exhibits a fair number of ironic utterances, mostly expressing emotions like disgust and anger. The experimental results show that the detection of some emotions (e.g., fear) is challenging compared to others and that the classifier suffers from data sparseness.

Pages: 33 to 37

Copyright: Copyright (c) IARIA, 2016

Publication date: November 13, 2016

Published in: conference

ISSN: 2519-8351

ISBN: 978-1-61208-519-7

Location: Barcelona, Spain

Dates: from November 13, 2016 to November 17, 2016