SC-Com: Spotting Collusive Community in Opinion Spam Detection

作者：

Highlights：

• Opinion spam aims to mislead potential customers by falsifying the genuine rating of products in online shopping platforms.

• We focus on their collusiveness to discriminate them from other genuine users and define collusiveness score which is used as a weight for projected user-user graph from review graph, from two primitive data such as rating and review time.

• After that, we detected community of collusive users and considered characteristics of each community and their role in each community as a feature for classification problem.

• By showing significant improvement of recall in actual collusive spam dataset, we demonstrated the effectiveness of our proposed model.

摘要

•Opinion spam aims to mislead potential customers by falsifying the genuine rating of products in online shopping platforms.•We focus on their collusiveness to discriminate them from other genuine users and define collusiveness score which is used as a weight for projected user-user graph from review graph, from two primitive data such as rating and review time.•After that, we detected community of collusive users and considered characteristics of each community and their role in each community as a feature for classification problem.•By showing significant improvement of recall in actual collusive spam dataset, we demonstrated the effectiveness of our proposed model.

论文关键词：Opinion spam detection,Clustering,Spammer,Social networks

论文评审过程：Received 6 October 2020, Revised 8 March 2021, Accepted 12 March 2021, Available online 26 March 2021, Version of Record 26 March 2021.

论文官网地址：https://doi.org/10.1016/j.ipm.2021.102593