Hello!
I recentlyfound an article that includes the following code to generate random walks using the node2vec algorithm.
MATCH (node)
CALL gds.alpha.randomWalk.stream('all
start: id(node),
steps: 15,
walks: 5,
// define randomwalk node2vec params
mode:'node2vec',
inOut:0.6,
return:1.0
})
YIELD nodeIds
// return name or title
RETURN [id in nodeIds |
coalesce(gds.util.asNode(id).name,
gds.util.asNode(id).title)] as walks
Unfortunately, I haven't been able to validate these settings in the GDS 2.3 library documentation.
Is there an alternate way of doing this? I don't want neo4j to generate node2vec embedding, just the random walks that I can use to feed into the Python implementation of node2vec.
For the record, I've read the adminitions against using the built-in neo4j node2vec method to generate embedding.
And, this is a newbie question: I've seen functions with blah.alpha.myfunc, blah.beta.myfunc, and blah.myfunc. While I understand alpha and beta software, I was hoping you could tell me the distinction in the context of neo4j.
Most cordially,
Jack Stewart
University of Michigan
mjstew@umich.edu
