NileULex: A Phrase and Word Level Sentiment Lexicon for Egyptian and Modern Standard Arabic

NileULex: A Phrase and Word Level Sentiment Lexicon for Egyptian and Modern Standard Arabic Samhaa R. El-Beltagy Nile University Juhayna Square, Sheikh Zayed City Giza, Egypt E-mail: samhaa@computer.org Abstract This paper presents NileULex, which is an Arabic sentiment lexicon containing close to six thousands Arabic words and compound phrases. Forty five percent of the terms and expressions in the lexicon are Egyptian or colloquial while fifty five percent are Modern Standard Arabic. The development of the presented lexicon has taken place over the past two years. While the collection of many of the terms included in the lexicon was done automatically, the actual addition of any term was done manually. One of the important criterions for adding terms to the lexicon, was that they be as unambiguous as possible. The result is a lexicon with a much higher quality than any translated variant or automatically constructed one. To demonstrate that a lexicon such as this can directly impact the task of sentiment analysis, a very basic machine learning based sentiment analyser that uses unigrams, bigrams, and lexicon based features was applied on two different Twitter datasets. The obtained results were compared to a baseline system that only uses unigrams and bigrams. The same lexicon based features were also generated using a publicly available translation of a popular sentiment lexicon. The experiments show that usage of the developed lexicon improves the results over both the baseline and the publicly available lexicon. Keywords: Arabic sentiment analysis, Arabic sentiment lexicons, Arabic idioms, Arabic opinion mining 1. Introduction Sentiment lexicons are an essential component of many sentiment analysis tools. Because of their importance, many sentiment lexicons for the English language, as well as for other languages, have appeared over the years. The most commonly used English lexicons include SentiWordNet (Baccianella, Esuli, and Sebastiani 2010), Ben Liu’s opinion lexicon (Liu 2010), the MPQA subjectivity lexicon (Wilson, Wiebe, and Hoffmann 2005), and the NRC lexicon (Kiritchenko, Zhu, and Mohammad 2014). Over the past five years, the topic of Arabic Sentiment analysis and opinion mining has been gaining momentum, especially since the use of the Arabic language has been increasing consistently over various social media platforms, amongst which are twitter and Facebook (Neal 2013)(“Facebook Statistics by Country” 2012)(Farid 2013). However, publicly available Arabic sentiment lexicons are scarce. In fact, to the knowledge of the author the only publicly available Arabic sentiment lexicon is a translated version of the NRC word emotion association lexicon (EmoLex) (Mohammad and Turney 2013). In addition, most attempts to build Arabic lexicons have focused primarily on lexicons for Modern standard Arabic (Abdul-Mageed and Diab 2012)(Badaro et al. 2014)(Mahyouba, Siddiquia, and Dahaba 2014) (the work presented in (Al-Sabbagh and Girju 2010) is a notable exception). With the exception of EmoLex, these lexicons also do not include compound phrases and idioms. When analysing social media for sentiment, this can affect the results adversely since the language used on social media platforms, is predominantly colloquial Arabic where sentiment is sometimes expressed entirely using compound phrases or …

Content truncated to 500 words