Node property prediction

Node property prediction can be used for predicting either discrete labels or numerical values for nodes with supervised machine learning. The Neo4j Graph Data Science library offers node property prediction both as standalone algorithms and as pipelines for end-to-end workflows.

Algorithms

The Neo4j Graph Data Science library contains the following node property prediction algorithms:

Aura Graph Analytics
- GraphSAGE supervised

Pipelines

The Neo4j Graph Data Science library supports and end-to-end workflow for node property prediction through the following pipelines:

Beta
- Node classification pipelines
Alpha
- Node regression pipelines