Research map / Graph learning / Graph self-supervised learning
On Class Distributions Induced by Nearest Neighbor Graphs for Node Classification of Tabular Data
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add an artificial k-NN graph and use a graph network instead of an MLP | sDGN · 11 tabular and citation datasets | F1 | No clear effect |
| add k-NN graph with GCN | GCN · eGrid | F1: 0.95 -> 0.83 | Hurt |
| add k-NN graph with GCN | GCN · Occupancy | F1: 0.99 -> 0.90 | Hurt |
| add k-NN graph with GCN | GCN · DryBean | F1: 0.78 -> 0.71 | Hurt |
| add k-NN graph with GIN | GIN · Musk | F1: 0.99 -> 0.94 | Hurt |
| add k-NN graph with sDGN | sDGN · Adult | F1: 0.70 -> 0.72 | No clear effect |
| add k-NN graph with GCN | GCN · Citeseer | F1: 0.69 -> 0.71 | No clear effect |
| build k-NN graph with cosine similarity instead of Euclidean distance (ablation) | sDGN · eGrid | F1: 0.95 -> 0.84 | Hurt |
| build k-NN graph with cosine similarity instead of Euclidean distance (ablation) | GCN · Adult | F1: 0.70 -> 0.53 | Hurt |
| build k-NN graph with cosine similarity instead of Euclidean distance (ablation) | sDGN · Pubmed | F1: 0.87 -> 0.84 | Hurt |
| use cosine similarity k-NN graph with GIN (ablation) | GIN · DryBean | F1: 0.75 -> 0.79 | Helped |
| use more than one graph convolutional layer (ablation) | DGN model selection · validation sets | validation accuracy | No clear effect |
Closest papers by meaning
- When Do Graph Neural Networks Help with Node Classification? Investigating the Homophily Principle on Node Distinguishability (NeurIPS 2023 · similarity 0.88)
- A Theory-Based Evaluation of Nearest Neighbor Models Put Into Practice (NeurIPS 2018 · similarity 0.88)
- A Fair Comparison of Graph Neural Networks for Graph Classification (ICLR 2020 · similarity 0.88)
- Improving Distinguishability of Class for Graph Neural Networks (AAAI 2024 · similarity 0.87)
- Revisiting Nearest Neighbor for Tabular Data: A Deep Tabular Baseline Two Decades Later (ICLR 2025 · similarity 0.86)
- Graph Classification via Reference Distribution Learning: Theory and Practice (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.