Sentiment analysis (SA) that is also referred to as opinion mining (OM) is the process in which the thoughts, feelings, emotions, and perspectives of individuals about a certain product, item or organization on web-based applications like Twitter, Facebook, Instagram, blogs etc. is gathered and analyzed. This article discusses a comprehensive examination of SA and its levels. The major focus of this manuscript is on aspect-based SA, as it aids manufacturing companies, to make better decisions by analyzing the perspectives and views of people about their products. The review discusses the various methodologies and techniques associated with the Aspect Based Sentiment Analysis (ABSA). In traditional methods, the features related to the aspects were drawn out manually, which makes it a time-consuming and error-prone process. Nevertheless, with the advancement of artificial intelligence, these limitations can be overpowered. Therefore, researchers nowadays, are employing artificial intelligence-based machine learning (ML) and deep learning (DL) techniques for enhancing the efficacy of ABSA. The automated general procedure for determining aspects from texts in the light of AI is also delineated in this manuscript. In addition to this, some of the recently published ABSA methods based on ML and DL are reviewed and compared and based on this review research gaps found in both techniques are also mentioned and highlighted.
Key words: Sentiment analysis, Aspect based sentiment analysis, Machine learning, Deep learning algorithms in SA.