Graph Classification on TUDataset Suite
78.6NCI1 AccuracyRényiCL
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| RényiCLEvaluation protocol=linear evaluation, Backbone=Graph isomorphism network (GIN), Average runs=52022.08 | 78.6 | 75.11 | 78.98 | 90.22 | 71.88 | 90.92 | 56.18 | 72.38 | |
| JOAOv2Evaluation protocol=linear evaluation, Backbone=Graph isomorphism network (GIN), Average runs=52022.08 | 78.36 | 74.07 | 77.4 | 87.67 | 69.33 | 86.42 | 56.03 | 70.83 | |
| JOAOEvaluation protocol=linear evaluation, Backbone=Graph isomorphism network (GIN), Average runs=52022.08 | 78.07 | 74.55 | 77.32 | 87.35 | 69.5 | 85.29 | 55.74 | 70.21 | |
| GraphCLEvaluation protocol=linear evaluation, Backbone=Graph isomorphism network (GIN), Average runs=52022.08 | 77.87 | 74.39 | 78.62 | 86.8 | 71.36 | 89.53 | 55.99 | 71.14 | |
| InfoGraphEvaluation protocol=linear evaluation, Backbone=Graph isomorphism network (GIN), Average runs=52022.08 | 76.2 | 74.44 | 72.85 | 89.01 | 70.65 | 82.5 | 53.46 | 73.03 |