Preferential attachment

In some cases, it is legitimate to assume the following:

More popular people are more likely to be linked to other popular people.

In these cases, the preferential attachment method has to be used. The preferential attachment score of nodes u and v is equal to the product of the degrees of nodes u and v:

score(u, v) = | N(u) | × | N(v) |

This score is also implemented in the GDS plugin under the gds.alpha.preferentialAttachment function:

MATCH (u), MATCH (v)
RETURN  id(u), id(v), 
        gds.alpha.linkprediction.preferentialAttachment(u, v, {
                        relationshipType: "REL", 
                        direction: "BOTH"
})

Note that we can optionally specify the type and direction of the relationships we use to find the neighborhood. We will look at an example application in the following section.