Abstract
Aspect-based sentiment analysis (ABSA) is a natural language processingproblem that requires analyzing user-generated reviews to determine: a) Thetarget entity being reviewed, b) The high-level aspect to which it belongs, andc) The sentiment expressed toward the targets and the aspects. Numerous yetscattered corpora for ABSA make it difficult for researchers to identifycorpora best suited for a specific ABSA subtask quickly. This study aims topresent a database of corpora that can be used to train and assess autonomousABSA systems. Additionally, we provide an overview of the major corpora forABSA and its subtasks and highlight several features that researchers shouldconsider when selecting a corpus. Finally, we discuss the advantages anddisadvantages of current collection approaches and make recommendations forfuture corpora creation. This survey examines 65 publicly available ABSAdatasets covering over 25 domains, including 45 English and 20 other languagesdatasets.