Sentiment analysis of Arabic students feedback throughout the semester

Nowadays, sentiment analysis has become a significant area of research related to machine learning and natural language processing, where sentiments and opinions about several topics are available on the Internet. This allows many parties such as companies, customers or governments to gauge and assess thousands of online opinions automatically. Sentiment on Arabic text is still some way behind that of English, mainly due to the inherent challenges in the language. We apply sentiment analysis to higher level education, by taking feedback from Arabic-speaking students every week during the semester in the form of 'Tweet'-like comments. This also has the benefit of allowing the students to express their opinion in their mother language to obtain a good quality feedback and alert the lecturer to any issues. The research shows that using sentiment analysis gives realistic sentiment values for the students' feedback, mainly due to the high accuracy of the Arabic sentiment lexicon. Over time the sentiment fluctuated in a plausible manner given the order of content presented.