A framework for sentiment analysis in Arabic text

Abstract

Over the last decade there has been an increase in number of E-mails or comments to a company via social media sites, to satisfy their customers, the company must take in to consideration these messages and comments and know whether the customers are satisfied with what the company offers or not. Several techniques have been proposed to analyze the sentiment of the comment writer. Dealing with the Arabic language is faced with many challenges, such as it is a morphologically rich language and how to return the word to its original root. In this paper the challenges of dealing with the Arabic language were reviewed and a framework was also established to analyze the comments in Arabic and classify it into positive, negative or neutral sentiment. The framework was trained and tested and then the conclusions were drawn based on its work.