Paper Details
Online Product Review sentiment analysis using Machine Learning Approaches
Authors
Gangavaram Riyaz, Thudumala Padmaja, Guddeti Venkata Nagi Reddy, Sampangi Venkata Ramesh, Nasanakota Rameshbabu, V.Silpa
Abstract
Reviews are a common kind of opinion spam in online markets. People are frequently employed to write extremely good or bad reviews for particular brands in an effort to promote or hinder them. This is frequently carried out in groups. Few studies have looked into identifying and analysing opinion spam groups that target a brand as a whole rather than just its products, despite some earlier attempts to do so. In this post, a set of 923 possible reviewer groups was manually classified using the reviews we gathered from the Amazon product review website. In order to cluster users who have mutually reviewed a large number of brands' products together, frequent item set mining over brand similarities is used to extract the groups. We postulate that eight characteristics unique to a (group, brand) pair determine the makeup of the reviewer groups. To categorize prospective groups as extremist entities, we create a supervised model based on features. For the job of classifying a group based on user ratings, we run various classifiers to see if the group exhibits extremity manifestations. The best classifier is found to be a three-layer perceptron-based classifier. To better understand the dynamics of brand-level opinion fraud, we continue to examine the actions of these groups in greater detail. Consistency in ratings, review sentiment, confirmed purchases, review dates, and helpful votes earned on reviews are examples of these characteristics. Surprisingly, we find that many certified reviewers exhibit strong sentiment, which, upon closer examination, reveals ways to get beyond the current safeguards against unauthorized incentives on Amazon.
Keywords
Online Product, Machine Learning, Naïve Bayes, Websites, Sentiment Analysis.
Citation
Online Product Review sentiment analysis using Machine Learning Approaches. Gangavaram Riyaz, Thudumala Padmaja, Guddeti Venkata Nagi Reddy, Sampangi Venkata Ramesh, Nasanakota Rameshbabu, V.Silpa. 2025. IJIRCT, Volume 11, Issue 2. Pages 1-5. https://www.ijirct.org/viewPaper.php?paperId=2504013