Sentiment Analysis of Amazon Product Reviews Using Hybrid Rule-Based Approach

In the past few years, retail market industries have taken a broad form to sell the products online and also to give the opportunity to customers to provide their valuable feedbacks, suggestions and recommendations. The aim of this paper is to provide an automatic comment analyzer. And propose an automatic comment analyzer and classification system which can determine the polarity of the customer comments collected from Amazon and Flipkart data domains effectively. This system should be able to process the large number of reviews. It should categorize the comments as positive, negative and neutral classes using five prime supervised learning classifiers such as NB, LR, SentiWordNet, RF and KNN. The paper also discusses their experimental results and challenges found. Therefore, this study shows the maximum usage of feature extraction, positive-negative sentiment, Amazon web source, mobile phone for a large set of reviews in the existing algorithms. It included the preliminary definitions, information extraction and retrieval aspects, role of machine learning, and the comment mining. The comment analysis and classification described comment polarity, orientation, subjectivity detection, comment analysis, summarization and classification. The classification of comment analysis techniques explained various lexicons and supervised algorithms along with the essential used-based concerns.