Systematic literature review of arabic aspect-based sentiment analysis
Salha Alyami, Areej Alhothali, Amani Jamal
Abstract
Aspect-based sentiment analysis (ABSA) is a natural language processing task that provides a detailed analysis of clients’ opinions about various aspects of a product or service. Although several review papers have examined Arabic ABSA studies, the number of studies covered is small, or the included studies are inadequately analyzed. Moreover, only one systematic literature review devoted to Arabic ABSA has been published to our knowledge, which covered only 21 primary studies. Therefore, this systemic literature review was conducted to analyze the existing techniques and resources used for Arabic ABSA. This review covered 47 primary studies published between 2012 and 2021 that were retrieved from eight bibliographic databases and search engines. The included studies were analyzed according to the dataset utilized, domain covered, Arabic language type, preprocessing procedures, selected features, word representation, employed techniques, and evaluation metrics used to assess the proposed techniques. As a result of this analysis, different limitations and issues were identified, and multiple future research directions were suggested. A new taxonomy was also constructed for the techniques employed, which were classified according to aspect-based sentiment analysis tasks and approaches.