Node Similarity Algorithm with Regex's

Neo4j Graph Platform Graph Algorithms/Graph Data Science
1

Hello,

How can I use the node similarity algorithm to determine how nodes are similar based on a property of a node? Would I, or can I, implement regular expressions here? Let's say I have a node label as BusinessLaw, and there is a property titled type. Can I use the node similarity algorithm to find the BusinessLaw nodes that are similar to other BusinessLaw nodes based on the same type property or a similar name for the type property?

I have a type property on BusinessLaw that says "BuildingProperty" and another type that says "CommercialProperty." How can I implement this algorithm to find nodes that are similar based on the property node value?

Thank you.

2

I think you're looking for the text similarity functions in APOC: Text Functions - APOC Extended Documentation


Node Similarity, in GDS, is built to work on neighboring nodes (eg. how many neighbors do these nodes have in common) - not their properties - and our similarity functions that use properties (cosine, jaccard, etc) use numerical values, not strings.