A Systematic Review of Sentiment Analysis Techniques, Challenges and Future Directions for Arabic Dialects

Abstract

At present time, the internet and social media are crucial for content sharing and expressing opinions on topics. These opinions can be analysed to assess product quality, identify issues, or improve products, but manual analysis of thousands of comments can be time-consuming. Therefore, Sentiment analysis is a method that uses natural language processing and computational linguistics to identify and extract subjective information from texts, determining the emotional state expressed, whether positive, negative, or neutral. A systematic literature review (SLR) on sentiment analysis of Arabic dialects (SAAD) is presented in this study. The main causes of diversity among these dialects are variations in syntax, lexicon, and grammar, which makes it challenging for scholars to classify DA polarity. This study has determined every stage that significantly affects the machine learning model used for dialect sentiment analysis, including text pre-processing, text annotation, feature extraction, and the approaches used. Additionally, the study has identified the issues and unresolved problems with sentiment analysis of the Arabic dialect (SAAD), which should be the main focus of future studies.