Sentiment Analysis of Arabic Dialects: A Review Study

Abstract

Arabic people use Arabic dialects on social media platforms to express their opinions and connect. Due to the absence of standard rules or grammar, Arabic dialects are more challenging for NLP tools to analyze than standard Arabic. While most review studies in this field have focused on highly indexed databases such as Scopus, Web of Science, and IEEE, these databases are not accessible to many Arabic researchers in Arabic countries due to financial constraints. This review study explores recent research and studies published in different databases to address this gap. The study identifies the most common sentiment analysis approaches, preprocessing and feature extraction techniques, and classification and evaluation techniques used in this field. The authors found that Twitter is the most commonly utilized source for researchers to collect their datasets, and machine learning approaches are the most commonly used for sentiment analysis in Arabic dialects. Overall, this study provides valuable insights into the challenges and opportunities for sentiment analysis in Arabic dialects.