A study on review manipulation classification using decision tree

Abstract

Identifying review manipulation has become one of hot research issues in e-commerce because more and more customers make their purchase decisions based on some personal comments from virtual communities and e-business websites. Customers consider these personal reviews are more reliable than the existing internet advertisements. Consequently, some enterprises attempt to create fake personal comments to affect customer behaviors and increase their sales. But, how to identify those manipulated reviews is a difficult task for customers. Therefore, this study employs Decision Tree (DT) to improve the classification performance of review manipulation by introducing eight potential review manipulation attributes. In addition, we attempted to discover the important factors of identifying manipulated reviews using correlation analysis and extracted knowledge rules. Finally, a real case of online users' comments regarding smart phones has been employed to testify the effectiveness of the proposed method.