Sentiment Analysis on Amazon Product Reviews using LSTM and Naive Bayes
Marella Sai Meghana, Davuluru Abhijith, Shaik Aysha, Praveen Kumar Kollu · 2023
The majority of consumers who prefer to purchase goods online from e-commerce websites frequently rely on reviews that have previously been written by other consumers or a summary of those reviews. For manufacturers and enterprises, opinion data is crucial. To systematically look through every post on the internet and glean insightful viewpoint information from it would be unrealistic. There is too much data to handle if you do it manually. Sentiment analysis is particularly utilized to find favorable, unfavorable, or inconsistent features of products in online text evaluations. The previous studies have given reviews only for a particular category of products. But the current technologies have failed to give the overall review of each product. This approach is used to provide the ultimate review of each online product, whether it is good, negative, or neutral, to overcome this drawback. To conduct the study between two different supervised learning techniques, Long short-term memory and Naive Bayes, have been attempted on online Amazon products. Their accuracies have then be compared.