DUTH at SemEval-2018 Task 2: Emoji Prediction in Tweets

Proceedings of the 12th International Workshop on Semantic Evaluation (SemEval-2018), pages 466–469 New Orleans, Louisiana, June 5–6, 2018. ©2018 Association for Computational Linguistics DUTH at SemEval-2018 Task 2: Emoji Prediction in Tweets Dimitrios Effrosynidis Georgios Peikos Symeon Symeonidis Avi Arampatzis Database & Information Retrieval research unit, Department of Electrical & Computer Engineering, Democritus University of Thrace, Xanthi 67100, Greece {deffrosy,georpeik,ssymeoni,avi}@ee.duth.gr Abstract This paper describes the approach that was de- veloped for SemEval 2018 Task 2 (Multilin- gual Emoji Prediction) by the DUTH Team. First, we employed a combination of pre- processing techniques to reduce the noise of tweets and produce a number of features. Then, we built several N-grams, to represent the combination of word and emojis. Finally, we trained our system with a tuned LinearSVC classifier. Our approach in the leaderboard ranked 18th amongst 48 teams. 1 Introduction Emojis are used in everyday life to express words or feelings of microblogging users. They are com- monly placed at the end of a sentence or alone. In this paper, we show how our emoji prediction framework was applied to SemEval-2018 Task 2 (Multilingual Emoji Prediction) (Barbieri et al., 2018), specifically on Subtask 1 (Emoji Prediction in English). In the last few years, many studies concentrated on emoji prediction and analysis. The prediction of emojis, the connection of emojis and words, and their separation from content-based tweet mes- sages, based on Long ShortTerm Memory net- works (LSTMs), was examined by Barbieri et al. (2017). The combination of emojis and sentiment was investigated by Novak et al. (2015), who de- veloped the first emoji sentiment lexicon and cre- ated a sentiment map of the 751 most frequently used emojis. The study of Barbieri et al. (2016) tested several skip-gram word embedding mod- els to measure the difference in performance be- tween machine-learning models and human anno- tation. Na’aman et al. (2017) analyzed the viabil- ity of a trained classifier to differentiate between those emojis utilized as semantic substance words and those utilized as paralinguistic or emotional multimodal markers. Miller et al. (2017) investi- gated the hypothesis of previous works that emo- jis in their regular textual contexts would gener- ously reduce and lead to miscommunication, but they found no such evidence; the potential for mis- communication appeared to be the same. The rest of this paper is organized as follows. Section 2 describes the architecture of our system and the dataset. In Section 3, we discuss the vari- ous parameters that were used to fine-tune the sys- tem, and present the performance of our frame- work. In Section 4, we lay out our main conclu- sions and research issues for further investigation. 2 System Description The principal goal of SemEval-2018 Task 2 - Sub- task 1 was emoji prediction in English. The frame- work we utilized consists of a bag of-words rep- resentation and N-gram extraction. We used the popular machine learning tool for Python, called Scikit-Learn (Pedregosa et al., 2011). 2.1 Preprocessing For the preprocessing of tweets, we were guided by …

Content truncated to 500 words