You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It would be great to see the PageRank implementation expanded to allow for Personalized PageRank (PPR) calculations.
Neo4j implements this feature as the simple argument sourceNodes; allowing a list of nodes/nodeIDs as the source possibilities (see table 2).
Another approach toward implementing this would be similar to how Networkx allows input of a personalization vector, which can contain initial starting probabilities for each node. PPR can therefore be performed by constraining all nodes except the wanted source nodes' probabilities to zero.
The text was updated successfully, but these errors were encountered:
It would be great to see the PageRank implementation expanded to allow for Personalized PageRank (PPR) calculations.
Neo4j implements this feature as the simple argument
sourceNodes
; allowing a list of nodes/nodeIDs as the source possibilities (see table 2).Another approach toward implementing this would be similar to how Networkx allows input of a
personalization vector
, which can contain initial starting probabilities for each node. PPR can therefore be performed by constraining all nodes except the wanted source nodes' probabilities to zero.The text was updated successfully, but these errors were encountered: