Abstract

In the current scenario, Internet registration is growing rapidly. Social media generates many signups daily with customer reviews, comments, and reviews. This huge amount of user-generated data is useless unless some mining is applied. Because there are so many fake reviews, review mining techniques should include spam detection to provide an authoritative review. Nowadays, some people use social media reviews to name themselves in purchasing products or services. Opinion spam is difficult to detect because many fake or fake comments have been created by groups or by humans for various purposes. They write fake reviews to mislead readers or sell automated detection devices to targeted products or to sell or downgrade them to tarnish their image. Proposed approaches include ontology, geographic region, IP vs. tracking, a dictionary of spam words using Naïve Bayes, simple brand assessment detection, and tracking account.

Description