Hierarchy Induction on Multi-source research dataset publications, patents, and funding records
67.7PrecisionTERMNET
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| TERMNETInput representation=extracted terms2026.06 | 67.7 | 99.22 | 36.7 | 94.44 | 4.5 | 4.5 | 5 | |
| TERMNETInput representation=raw text2026.06 | 66.39 | 99.57 | 33.11 | 69.94 | 4 | 2.5 | 5 | |
| K-MeansInput representation=terms2026.06 | 61.12 | 99.75 | 30 | 86.42 | 3 | 1.5 | 4 | |
| SCYCHIC2026.06 | 60.52 | 99.73 | 23.48 | 79.02 | 4 | 3 | 4 | |
| K-MeansInput representation=raw text2026.06 | 53.28 | 99.79 | 24.5 | 84.3 | 2 | 2.5 | 2.5 | |
| K-MeansInput representation=summary2026.06 | 47.56 | 99.41 | 23.98 | 91.25 | 3.5 | 2 | 4 |