NODES 2026: Graph-Aware Recommendations with Neo4j GDS and Random Forest
A full technical walkthrough of the NODES 2026 talk: bipartite knowledge graph design, FastRP embeddings, graph-aware hard negative sampling via constrained random walks, stratified resampling to correct power-law skew, Random Forest on element-wise embedding products (OOB AUPRC 97.3%), and Platt re-calibration from 50% sampled prior to 5.5% population prior.
Read the post