Multi-label Text Classification on Ohsumed
72.6Micro-F1S2TC-BDD
Evaluation results
| Method | Links | ||
|---|---|---|---|
| S2TC-BDDNumber of labeled texts (Nl)=10,0002026.03 | 72.6 | 67.8 | |
| BERT+AMNumber of labeled texts (Nl)=10,0002026.03 | 72.1 | 66.6 | |
| BERT+BCENumber of labeled texts (Nl)=10,0002026.03 | 69.5 | 65.7 | |
| CAPNumber of labeled texts (Nl)=10,0002026.03 | 64.5 | 61.5 | |
| MetaExpertNumber of labeled texts (Nl)=10,0002026.03 | 64.4 | 61.1 | |
| S2TC-BDDNumber of labeled texts (Nl)=1,0002026.03 | 63.6 | 56.2 | |
| BERT+AMNumber of labeled texts (Nl)=1,0002026.03 | 60.4 | 55.2 | |
| BERT+BCENumber of labeled texts (Nl)=1,0002026.03 | 58.3 | 51.2 | |
| CAPNumber of labeled texts (Nl)=1,0002026.03 | 57.1 | 51.2 | |
| MetaExpertNumber of labeled texts (Nl)=1,0002026.03 | 55.8 | 51.2 | |
| S2TC-BDDNumber of labeled texts (Nl)=1002026.03 | 44.7 | 30.8 | |
| SDRLNumber of labeled texts (Nl)=10,0002026.03 | 44.3 | 21.9 | |
| BERT+AMNumber of labeled texts (Nl)=1002026.03 | 42.1 | 29.8 | |
| MetaExpertNumber of labeled texts (Nl)=1002026.03 | 41.8 | 30 | |
| SDRLNumber of labeled texts (Nl)=1,0002026.03 | 39.7 | 19 | |
| CAPNumber of labeled texts (Nl)=1002026.03 | 38 | 25.2 | |
| SDRLNumber of labeled texts (Nl)=1002026.03 | 31.1 | 10.7 | |
| BERT+BCENumber of labeled texts (Nl)=1002026.03 | 27.8 | 13.1 |