A prediction method enhanced by the degree of nodes

Many link prediction methods via local or global information have been proposed, which are important measures to infer relationship between members in an incomplete network. Given an incomplete network, the common methods we adopted are some similarity-based measures, which still have defects and imperfect parts. In order to enhance the prediction result, we propose a method based on the degree of nodes. Experiments show the approach discards some drawbacks of single similarity methods and AUC is higher in real-world networks, especially in those with fewer communities.