The paper provides a framework for obtaining biased node samples by crawling the network under supervision.
The paper introduces a new supervised sampling problem on networked data.
The proposed method is to sample a sub network under a specific goal of, acquiring positive nodes.