Skip to main content

Table 3 Baseline methods for node pair (u, v) with neighbour sets N(u) and N(v). \(\hat {N}\)(x) are the neighbours of the neighbours of x

From: Neural networks for link prediction in realistic biomedical graphs: a multi-dimensional evaluation of graph embedding-based approaches

Name Definition definition
Adamic-Adar \(\frac {1}{log(|N(u) \cap N(v)|)}\) \(\frac {1}{log(|N(u) \cap \hat {N}(v)|)}\)
Common Neighbours |N(u)∩N(v)| \(|N(u) \cap \hat {N}(v)|\)
Jaccard Index \(\frac {|N(u) \cap N(v)|}{|N(u) \cup N(v)|}\) \(\frac {|N(u) \cap \hat {N}(v)|}{|N(u) \cup \hat {N}(v)|}\)