Abstract In this paper, we propose a semi-supervised approach for sentiment analysis of Arabic and
Arabic is one of the most morphologically complex languages, and its numerous dialects rend