A review of sentiment analysis research in Arabic language
An in-depth qualitative study of the most important research works in this context by discussing strengths and limitations of existing approaches, and surveys both approaches that leverage machine translation or transfer learning to adapt English resources to Arabic and approaches that stem directly from the Arabic language.
Abstract
Sentiment analysis is a task of natural language processing which has\nrecently attracted increasing attention. However, sentiment analysis research\nhas mainly been carried out for the English language. Although Arabic is\nramping up as one of the most used languages on the Internet, only a few\nstudies have focused on Arabic sentiment analysis so far. In this paper, we\ncarry out an in-depth qualitative study of the most important research works in\nthis context by presenting limits and strengths of existing approaches. In\nparticular, we survey both approaches that leverage machine translation or\ntransfer learning to adapt English resources to Arabic and approaches that stem\ndirectly from the Arabic language.\n