amaddahi/ml_with_gds

Marvel Universe demo

Machine Learning with Neo4j's Graph Data Science Library

This library contains a demo of Neo4j's new node classification and link prediction algorithms - as applied to the Marvel Universe. It's built on top of Tomasnojo's Marvel Graph Gist available here.

This demo shows how to:

1. Prepare data for using an ML model
2. Train graph embeddings as an input for a model
3. Predict missing node labels (in this case, who's an X-Man)
4. Predict missing relationships (who will appear in a future comic book together)
5. Use the trained models to make predictions

I've also uploaded a bloom perspective, to let you explore your results visually.

For further documentation on using the GDS ML functions, check out our docs here

BrowserGuides

Added browser guide to allow to repeat/share the example

Can either:

1. Run from existing .html files in browserguide hosted under a web server
2. Build from adocs using render.sh