A Novel Defending Scheme for Graph-Based Classification Against Graph Structure Manipulating Attack.

Graph-based classification methods are widely used in social network. Wang et al. [1] proposed an attack for the collective classification method by manipulating the structure of the graph. In this paper, we propose a novel defense scheme for this attack by repairing the graph structure via deleting the edge for the key nodes. Our experiments demonstrate that our scheme is very effective in resisting such attacks and enables the classification algorithm to achieve pre-attack accuracy and precision.