DAICT: A Dialectal Arabic Irony Corpus Extracted from Twitter

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 6265–6271 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 6265 DAICT: A Dialectal Arabic Irony Corpus Extracted from Twitter Ines Abbes, Wajdi Zaghouani, Omaima El-Hardlo, Faten Ashour College of Humanities and Social Sciences, Hamad Bin Khalifa University Doha, Qatar {iabbesepmnasri, wzaghouani, oelhardlo}@hbku.edu.qa, fashour@mail.hbku.edu.qa Abstract Identifying irony in user-generated social media content has a wide range of applications; however to date Arabic content has received limited attention. To bridge this gap, this study builds a new open domain Arabic corpus annotated for irony detection. We query Twitter using irony-related hashtags to collect ironic messages, which are then manually annotated by two linguists according to our working definition of irony. Challenges which we have encountered during the annotation process reflect the inherent limitations of Twitter messages interpretation, as well as the complexity of Arabic and its dialects. Once published, our corpus will be a valuable free resource for developing open domain systems for automatic irony recognition in Arabic language and its dialects in social media text. Keywords: Arabic Dialects, Irony, Twitter, Corpus Generation 1. Introduction Irony detection has a broad range of applications in vari- ous research areas, including sentiment analysis, text min- ing and author profiling (Van Hee, 2017). Nowadays, social-media platforms allow users to express and share their opinions, attitudes and beliefs instantaneously. Users tend to employ figurative language devices, such as irony and sarcasm, to achieve different communication pur- poses (Hern´andez Far´ıas, 2017). This user-generated con- tent represents a rich source of valuable data. Handling this data is a challenging task in Natural Language Processing, particularly when dealing with ironic statements, with users intending to convey a message that goes beyond the literal meaning. Current scholarship is yet to reach an agreement on a uni- versal definition of the concept of irony (Van Hee, 2017; Hern´andez Far´ıas, 2017). Various definitions have been proposed (Grice et al., 1975; Wilson and Sperber, 1992; Kumon-Nakamura et al., 1995; Attardo, 2000; Giora et al., 2013) and irony is broadly defined as a rhetorical device used to express the opposite of what is literally meant (Grice et al., 1975). This study investigates the use of irony in Arabic Twitter messages. For the purpose of our analysis, we do not distinguish between irony and sarcasm, thereby using irony as an umbrella term. Annotated corpora for irony detection are now available in many languages; however, to the best of our knowledge, there is no publicly and freely available irony corpus for Arabic language or its dialects (Rosso et al., 2018; Za- ghouani, 2014). To bridge this gap, this research introduces a new irony corpus of Dialectal Arabic extracted from Twit- ter: DAICT1. This paper is structured as follows. The following section gives an overview of the existing work on irony corpora generation. Section 3 presents the data collection process as well as annotation guidelines and challenges. Section 4 elaborates the corpus analysis. We conclude this paper in 1The …

Content truncated to 500 words