SubClassOf triple classification on YAGO39K (test)
90AccuracyEIKE-PRE-EYE
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| EIKE-PRE-EYESampling strategy=uniform2024.01 | 90 | 86.56 | 94.7 | 90.45 | |
| EIKE-PRE-EYESampling strategy=bernoulli2024.01 | 89.05 | 85.6 | 93.9 | 89.56 | |
| TransEllipsoid*2024.01 | 85.1 | 82.1 | 89.7 | 85.7 | |
| TransFG*2024.01 | 84.5 | 78.6 | 95.2 | 86.1 | |
| TransC*2024.01 | 83.7 | 78.1 | 93.9 | 85.2 | |
| TransR*2024.01 | 80.4 | 74.7 | 91.9 | 82.4 | |
| TransH*2024.01 | 80.2 | 76.4 | 87.5 | 81.5 | |
| EIKE-UNP-MATSampling strategy=uniform2024.01 | 77.85 | 71.54 | 92.5 | 80.68 | |
| EIKE-UNP-MATSampling strategy=bernoulli2024.01 | 77.8 | 70.23 | 96.5 | 81.3 | |
| EIKE-PRE-MATSampling strategy=uniform2024.01 | 77.7 | 71.08 | 93.4 | 80.73 | |
| TransE*2024.01 | 77.6 | 72.2 | 89.8 | 80 | |
| EIKE-PRE-MATSampling strategy=bernoulli2024.01 | 77.2 | 70.15 | 94.7 | 80.6 | |
| EIKE-UNP-EYESampling strategy=bernoulli2024.01 | 76.85 | 69.28 | 96.6 | 80.67 | |
| EIKE-UNP-EYESampling strategy=uniform2024.01 | 76.8 | 68.77 | 98.2 | 80.89 | |
| TransD*2024.01 | 75.9 | 70.6 | 88.8 | 78.7 | |
| TransCuboid*2024.01 | 74.7 | 68.3 | 92.3 | 78.5 | |
| HolE*2024.01 | 70.5 | 73.9 | 63.3 | 68.2 | |
| DistMult*2024.01 | 61.9 | 68.7 | 43.7 | 53.4 | |
| ComplEx*2024.01 | 61.6 | 71.5 | 38.6 | 50.1 | |
| Concept2Vec2024.01 | 59.3 | 70.3 | 44.5 | 54.8 |