Archives

Fake Reviews Detection through Users Behavior Analysis


Ghaidaa A. Al-Sultany and Saba Mohammed Hussain
Abstract

Today, the world is moving towards technological development in all areas of life, especially in sales and purchase. Therefore, it is noted that many traders are trying to promote their goods by displaying them on many websites. Our model features additional analysis based on the analysis of user behavior in terms of calculating the similarity between previous comments and the number of comments per product by one user. The work also relied on the user's fairness account by calculating the rating. In this paper; User behavior has been analyzed to detect the customer's identity as fake or authentic. It depends on user history in terms of his past reviews and rating ratio. Using new attribute metrics derived from the user's own data set. The system is tested on the data collection collected by Amazon; two of these data sets were tested and the accuracy of the rating ranged from 80% to 95%.

Volume 11 | 10-Special Issue

Pages: 737-741

DOI: 10.5373/JARDCS/V11SP10/20192864