A Case Study on BERT models for Aspect Based Sentiment Analysis

Aspect Based Sentiment Analysis (ABSA) is a powerful tool to analyse a product and increase the business exponentially. The process involves classifying a sentiment to the text according to the different aspects such as positive or negative and sometimes as neutral. The deep learning algorithms (DLA) are more preferred for these tasks due to their better performance in getting more accurate and precise results as compared to the traditional machine learning algorithms used for the same tasks. The objective of this paper is to study the specific Bidirectional Encoder Representation (BERT) models used for ABSA tasks. The aim is to identify and analyse the various techniques and the architecture being used and also the challenges and issues associated with these models. An analysis is provided to summarize the recent implementation of the models and the different ABSA tasks performed by those models, the architecture and different techniques implemented with the models and the challenges and issues associated with the models.