User-based network embedding for opinion spammer detection

作者：

Highlights：

• This paper jointly learning direct relevance and indirect relevance for collective spammers detection.

• The direct relevance is captured by the pairwise behavior features between two users.

• The indirect relevance is learned based on a k-step co-rating neighborhood proximity in the network.

• Extensive experiments conducted on two real-world datasets verify the effectiveness of the proposed method.

摘要

•This paper jointly learning direct relevance and indirect relevance for collective spammers detection.•The direct relevance is captured by the pairwise behavior features between two users.•The indirect relevance is learned based on a k-step co-rating neighborhood proximity in the network.•Extensive experiments conducted on two real-world datasets verify the effectiveness of the proposed method.

论文关键词：Spam detection,Collective spammer,Network embedding,Signed network

论文评审过程：Received 18 March 2021, Revised 21 December 2021, Accepted 22 December 2021, Available online 29 December 2021, Version of Record 6 January 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108512