Tw-StAR at SemEval-2017 Task 4: Sentiment Classification of Arabic Tweets

Proceedings of the 11th International Workshop on Semantic Evaluations (SemEval-2017), pages 664–669, Vancouver, Canada, August 3 - 4, 2017. c⃝2017 Association for Computational Linguistics Tw-StAR at SemEval-2017 Task 4: Sentiment Classification of Arabic Tweets Hala Mulki 1, Hatem Haddad 2, Mourad Gridach 3 and Ismail Babaoglu 1 1 Department of Computer Engineering, Selcuk University, Konya, Turkey 2 Department of Computer and Decision Engineering, Universit´e Libre de Bruxelles, Belgium 3 High Institute of Technology, Ibn Zohr University, Agadir, Morocco halamulki@selcuk.edu.tr,Hatem.Haddad@ulb.ac.be m.gridach@uiz.ac.ma,ibabaoglu@selcuk.edu.tr Abstract In this paper, we present our contribu- tion in SemEval 2017 international work- shop. We have tackled task 4 entitled “Sentiment analysis in Twitter”, specifi- cally subtask 4A-Arabic. We propose two Arabic sentiment classification models im- plemented using supervised and unsuper- vised learning strategies. In both mod- els, Arabic tweets were preprocessed first then various schemes of bag-of-N-grams were extracted to be used as features. The final submission was selected upon the best performance achieved by the su- pervised learning-based model. Neverthe- less, the results obtained by the unsuper- vised learning-based model are considered promising and evolvable if more rich lex- ica are adopted in further work. 1 Introduction Social media is literally shaping decision making processes in many aspects of our daily lives. Ex- ploring online opinions is therefore becoming the focus of many analytical studies. Twitter is one of the most popular microblogging systems that enables a real-time tracking of opinions towards ongoing events (Saif et al., 2016). Hence, it pro- vides the needed feedback information for analyt- ical studies in several domains such as politics and targeted advertising (El-Makky et al., 2014). Sen- timent analysis plays an essential role in perform- ing such studies as it can extract the sentiments out of the opinions and classify them into polar- ities (Tang et al., 2015). Arabic language has re- cently been considered as one of the most grow- ing languages on Twitter with more than 10.8 mil- lion tweets per day (Alhumoud et al., 2015). Yet, Arabic is remarkably less tackled in the research of Sentiment Analysis (Nabil et al., 2015; ElSa- har and El-Beltagy, 2015). With more resources and tools for Arabic Natural Language Process- ing (NLP) becoming available, and with the re- cent developed sentiment lexica for Modern Stan- dard Arabic (MSA) and dialectal Arabic, this year, SemEval contest offers the opportunity to apply sentiment classification on Arabic tweets through subtask 4A-Arabic (Rosenthal et al., 2017). An- alyzing Arabic tweets is significantly challeng- ing due to the complex nature and morphology of the Arabic language. Furthermore, Arabic tweets are mostly informal and written in different di- alects in which same words or expressions may have drastically different sentiments. For exam- ple, ������� ����� is a compliment of a pos- itive sentiment that means “May GOD grant you health” in the Levantine dialect while it has an aggressive meaning of “burn in fire” in the Mo- roccan and Tunisian dialects (El-Makky et al., 2014). Additionally, tweeters tend to use abbrevi- ations, neologisms, emoji and sarcasm frequently …

Content truncated to 500 words