A Comprehensive Machine-Learning-Based Approach for Aspect-Based Sentiment Analysis Over Food Consumables

Due to the large growth of web technologies and online resources, there is high information made available to everyone. With the growth of technology and resources, people are expressing their opinions and thoughts more comfortably. It impacts the decision-making of everyone, the mindset of the customer is based on the reviews of the other users in the web resources. Hence, going through all the reviews and getting to a final conclusion about the product has become a major issue nowadays. The solution to the problem is provided by sentiment analysis. Sentiment analysis derives the final polarity of the product by analyzing all the reviews of users. The overall sentiment polarity is affected by aspects within the domain, knowing the polarity of every aspect in the considered domain can be done by ABSA (aspect-based sentiment analysis). Aspect-based sentiment analysis is developed using many different solutions. Among them, deep learning and machine-learning methods made a great impact on the development of ABSA. In ABSA, the primary feature to be taken into consideration is aspects of domain is polarity. Choosing the most suitable model from the different models which are available is the difficult part. A comparative analysis is performed between ML (machine learning) and DL (deep learning) models.