CENNLP at SemEval-2018 Task 2: Enhanced Distributed Representation of Text using Target Classes for Emoji Prediction Representation

Proceedings of the 12th International Workshop on Semantic Evaluation (SemEval-2018), pages 486–490 New Orleans, Louisiana, June 5–6, 2018. ©2018 Association for Computational Linguistics CENNLP at SemEval-2018 Task 2: Enhanced Distributed Representation of Text using Target Classes for Emoji Prediction Representation Hariharan V, Naveen J R, Barathi Ganesh H. B., Anand Kumar M, Soman K P Center for Computational Engineering and Networking (CEN) Amrita School of Engineering, Coimbatore Amrita Vishwa Vidyapeetham, India cb.en.p2cen16007@cb.students.amrita.edu, barathiganesh.hb@gmail.com , m anandkumar@cb.amrita.edu Abstract Emoji is one of the ”fastest growing language ” in pop-culture, especially in social media and it is very unlikely for its usage to decrease. These are generally used to bring an extra level of meaning to the texts, posted on social me- dia platforms. Providing such an added info, gives more insights to the plain text, arising to hidden interpretation within the text. This pa- per explains our analysis on Task 2, ” Multilin- gual Emoji Prediction” sharedtask conducted by Semeval-2018. In the task, a predicted emoji based on a piece of Twitter text are la- belled under 20 different classes (most com- monly used emojis) where these classes are learnt and further predicted are made for un- seen Twitter text. In this work, we have exper- imented and analysed emojis predicted based on Twitter text, as a classification problem where the entailing emoji is considered as a la- bel for every individual text data. We have im- plemented this using distributed representation of text through fastText. Also, we have made an effort to demonstrate how fastText frame- work can be useful in case of emoji prediction. This task is divided into two subtasks, they are based on dataset presented in two different lan- guages English and Spanish. 1 Introduction The consumption of technology in industry deliv- ers potential tools for communication. Messaging has turned into a critical method of communica- tion all through the world and is expanding at a quick rate. Adding emoji in the text convey lit- tle more information about the person’s emotion, which otherwise is absent in the normal text. Emo- tions contents of text is better expressed by usage of emojis. Emojis are fundamentally kind of im- age which are logically connected with the writ- ten text. Tweets and online social media platforms are investigated to assess the emotion depth of the several issues for sentiment analysis and Opinion mining in natural language platform. In recent times, the interest in theses area received is very much increased and made several classifications which are polarity based classification such as pos- itive, negative and neutral. In case of emojis cer- tain remarkable studies on emoji sematic and us- age find out in papers (Aoki and Uchida, 2011) Relevant study into emoji (Barbieri et al., 2018) are limited in number. The common exploration about emoji has inspired (Barbieri et al., 2017) on descriptive analysis or used them as a indication the emotional affect (Rathan et al., 2017) on so- cial media. That is too restricted in face emojis. 2 …

Content truncated to 500 words