A System for Rating and Reviewing Based on Features

Abstract

Online review sites are a key customer and user feedback source. However, there is rampant misuse of such online platforms either by corporate entities promoting their products through fake reviews to boost their scores or by competitors and disgruntled/troll users trying to tarnish the credibility of the product in question by review bombing the sites. This novel system is a reliable review/rating system that uses machine learning, NLP and sentimental analysis to deliver acute informative feature-based reviews by users that filter out the unnecessary reviews that undermine the process and produce a straightforward representation that is not ambiguous as well as being easy for the user to infer from the review. Thus, the system ensures a fair review process for both user-side and organizational purposes with minimal human resources requirements. The review moderation task is also significantly simplified so that only a few experienced moderator staff are needed to maintain the system more lenient to the time that requires their active attention.