A Comparative Review on Approaches of Aspect Level Sentiment Analysis

Abstract

Aspect-level sentiment analysis carries out a detail-oriented task of retrieving the aspects in a particular document and categorizing the opinion about every aspect. The applications of Aspect Level Sentiment Analysis (ALSA) are extended because more data is readily available on social networking sites. The applications include reviewing the opinions of customers, examining the mental health status of patients based on social media posts, market research and competitor analysis, product analysis, and stock price prediction. The extensive applications encouraged researchers to devise new techniques and approaches for ALSA. In this survey, the techniques and approaches of ALSA are reviewed. The review is carried out based on the categorization which classifies the studies into three approaches: Knowledge-based, Statistical, and Hybrid. Various techniques for implementing these approaches and performance on different datasets are reviewed. Furthermore, analyzing the benefits and drawbacks of each literature and also finds their research gaps to keep up with current trending research areas. This study is finally concluded by highlighting few research challenges and further research directions in this domain.