A Multilevel Scheme for Irony Annotation in Social Network Content
Jihen Karoui, Farah Benamara, Véronique Moriceau · 2019
Considering the work carried out in the field of linguistics on verbal irony markers in poems, novels, etc., this chapter shows that work on irony in social networks from a computational perspective has barely scratched the surface of the problem, without going into specifics. It proposes a first multilevel annotation scheme and a corpus of tweets annotated using this scheme. It first presents the study corpus, French Irony Corpus, followed by the annotation scheme. The chapter then describes those categories of irony used in the annotation approach alongside those proposed by linguists. It describes the procedure used to annotate tweets using the proposed scheme via the Glozz annotation tool, followed by a presentation of the quantitative and qualitative results of this campaign. Particular attention is given to the interactions between types of irony triggers and markers, irony categories and markers, and the impact of external knowledge on irony detection.