Chinese Journal of Network and Information Security ›› 2018, Vol. 4 ›› Issue (5): 39-46.doi: 10.11959/j.issn.2096-109x.2018042

• Papers • Previous Articles     Next Articles

Spammer detection technology of social network based on graph convolution network

Qiang QU,Hongtao YU,Ruiyang HUANG   

  1. National Digital Switching System Engineering &Technological R&D Center,Zhengzhou 450002,China
  • Revised:2018-05-05 Online:2018-05-15 Published:2018-08-04
  • Supported by:
    The National Natural Science Foundation Innovation Group Project(61521003)


In social networks,Spammer send advertisements that are useless to recipients without the recipient's permission,seriously threatening the information security of normal users and the credit system of social networking sites.In order to solve problems of extracting the shallow features and high computational complexity for the existing Spammer detection methods of social networks,a Spammer detection technology based on graph convolutional network(GCN) was proposed.Based on the network structure information,the method introduces the network representation learning algorithm to extract the network local structure feature,and combines the GCN algorithm under the re-regularization technology condition to obtain the network global structure feature to achieve the goal of detecting Spammer.Experiments are done on social network data of results show that this method has high accuracy and efficiency.

Key words: cyberspace security, Spammer detection, network representation learning, GCN

CLC Number: 

No Suggested Reading articles found!