Identifying the sentiment of particular features or aspects of a service or product is the goal of the NLP subfield known as Aspect-Based Sentiment Analysis (ABSA). ABSA is a challenging task that requires the analysis of multiple aspects in a single text, as well as the identification of the sentiment of each aspect. This study presents a comprehensive summary of ABSA, including its importance, challenges, applications, and recent advances. We review state-of-the-art ABSA models, datasets, and evaluation metrics and comprehensively review their advantages and drawbacks. We also discuss the current trends and future directions in ABSA research.

Sentiment Analysis, Aspect-Based Sentiment Analysis, Aspect Extraction, Sentiment classification.