Abstract
Mining publicly available data for meaning and value is an important research direction within social media analysis. Even for automatically analyzing collected textual data, a manual effort is needed for a successful machine learning algorithm to effectively classify text. Corpus annotation is labeling datasets with appropriate classes. There is a lack in the Arabic annotated corpus although Arabic is one of the languages that shows a fast uptake of sentiment analysis research, despite limited resources and scarce annotated corpora. In this paper, we review the most recent work on annotation carried out for papers focusing on Arabic sentiment analysis, between the years of 2010 and 2016.