Exploring Review Spammers by Review Similarity: A Case of Fake Review in Taiwan

Understanding the phenomenon of spam reviews in social media is now an emerging and important issue since some enterprises may hire spammers to post fake reviews to promote their product or demote product of their competitors. The hired spammers are paid based on the fake reviews. Thus, these spammers may rewrite previous reviews as new review to earn the money. Thus, review similarity maybe a cue to detect fake reviews. Although literature had investigated the spam reviews, the review similarity of real review spammers is relatively unexplored. The objective of this paper is to explore the review spammers with a real case of fake reviews in Taiwan by investigating the cosine similarity and content length of reviews. We have proposed a text mining approach for a better understanding the phenomenon of fake reviews. The empirical results suggested that when comparing with normal reviews, the spam reviews were longer and with higher content similarity.