Node Regression for nodes with properties and no relationships

Neo4j Graph Platform Graph Algorithms/Graph Data Science
1

Hello,

I'm using Neo4j Desktop v5.26.1

I'm not new to Neo4j or Cypher but I am new to the GDS and ML side of things, so please forgive newbie GDS/ML question.

I am trying to create a very simple Node Regression pipeline based on nodes with node properties and no relationships. Very simple ML problem to start my ML pipeline development.
AKA: gds.alpha.pipeline.nodeRegression

If I start off with 3 nodes, say:

create (:node{x:12, y:1})
create (:node{x:14, y:2})
create (:node{x:16, y:3})

I'm looking to create an ML pipeline which will produce a model with predictions similar to:
x = 10 + 2y

I understand I will need more training data than 3 nodes but this is just to display the point.

I have created a projection using 'node' and x and y properties with '*' as the relationship (projects no relationships as there aren't any in the graph)

I then create a Node Regression Pipeline
gds.alpha.pipeline.nodeRegression.create('test01")

My issue is what do I add into this pipeline in terms of addNodeProperty, selectFeatures and addLinearRegression to achieve a simple ML algorithm based on the x and y properties.

All the examples I have looked at relate to relationships. Is this just too simple a ML task for GDS to work with? Is this possible?

Thank you community.