Node similarity and degree cutoff

Neo4j Graph Platform Graph Data Science / Graph Analytics
1

Hi!
The nodeSimilarity algorithm works with two sets of nodes A, and B where all the relationships goes from A to B.
The algorithm accepts a parameter, degreeCutoff described as:
"Inclusive lower bound on the node degree for a node to be considered in the comparisons. This value can not be lower than 1."

However, this only seem to apply to nodes in the set A, not in B, which is what I would want.
Is this a bug or working as intended?

Thanks

2

actually i want to use the upperDegreeCutoff, but that also seem to apply to set A only..

3

Its working as intended, for the target node we dont have a cutoff option (yet).
You could use filtered NodeSimilarity and provide a targetNodeFilter which only includes these nodes.
But with larger graphs the list could grow quite big if you refer to them by id.
So, you could use gds.graph.nodeLabel.mutate to pass the filter as a label instead of a list.