SINAI at SemEval-2019 Task 3: Using affective features for emotion classification in textual conversations

Proceedings of the 13th International Workshop on Semantic Evaluation (SemEval-2019), pages 307–311 Minneapolis, Minnesota, USA, June 6–7, 2019. ©2019 Association for Computational Linguistics 307 SINAI at SemEval-2019 Task 3: Using affective features for emotion classification in textual conversations Flor Miriam Plaza-del-Arco, M. Dolores Molina-Gonz´alez, M. Teresa Mart´ın-Valdivia, L. Alfonso Ure˜na-L´opez Department of Computer Science, Advanced Studies Center in ICT (CEATIC) Universidad de Ja´en, Campus Las Lagunillas, 23071, Ja´en, Spain {fmplaza, mdmolina, maite, laurena}@ujaen.es Abstract Detecting emotions in textual conversation is a challenging problem in absence of nonverbal cues typically associated with emotion, like fa- cial expression or voice modulations. How- ever, more and more users are using message platforms such as WhatsApp or telegram. For this reason, it is important to develop systems capable of understanding human emotions in textual conversations. In this paper, we carried out different systems to analyze the emotions of textual dialogue from SemEval-2019 Task 3: EmoContext for English language. Our main contribution is the integration of emo- tional and sentimental features in the classifi- cation using the SVM algorithm. 1 Introduction Emotions seem to govern our daily lives since most of our decisions are guided by our mood. They are complex and that is why they have been studied in many areas over time. Given the im- portance to develop systems to be able to mimic functioning of the human brain, emotions have at- tracted the attention in the field of affective com- puting (Thilmany, 2007). To our knowledge, there are not many works that focus on studying how emotions are reflected verbally. However, studying emotions on text messaging platforms such as WhatsApp, Face- book Messenger or Telegram is important as more and more users are using them to share their expe- riences and emotions. Currently, detecting emotions in instant mes- saging has multiple applications in different fields (Gupta et al., 2017; Yadollahi et al., 2017; Hakak et al., 2017), such as businesses intelligence to in- crease customer satisfaction knowing their prefer- ences, social media to alert users if they are going to post an offensive tweet or psychology to detect some disorders like anorexia, anxiety or stress. In this paper, we present the different sys- tems we developed as part of our participation in SemEval-2019 Task 3: Contextual Emotion De- tection in Text (EmoContext) (Chatterjee et al., 2019b). It is an emotion classification task. Given a textual dialogue along with two turns of context, its consists of classify the emotion of user utter- ance as one of the emotion classes: Happy, Sad, Angry or Others. The rest of the paper is structured as follows. In Section 2 we explain the data used in our methods. Section 3 introduces the lexical resources used for this work. Section 4 presents the details of the pro- posed systems. In Section 5, we discuss the analy- sis and evaluation results for our system. We con- clude in Section 6 with remarks and future work. 2 Data To run our experiments, we used the English datasets provided by the …

Content truncated to 500 words