Aspect-level sentiment analysis on Rest 14
90.86AccuracyLSAT
Evaluation results
| Method | Links | |
|---|---|---|
| LSAT2023.02 | 90.86 | |
| ABSA-DeBERTa2023.02 | 89.46 | |
| HGCN+BERTBackbone=BERT, Evaluation Protocol=Best performance2022.04 | 87.41 | |
| HGCN+BERTBackbone=BERT, Evaluation Protocol=Average performance2022.04 | 86.45 | |
| DualGCN+BERTBackbone=BERT2022.04 | 86.29 | |
| InstructABSAvariant=12023.02 | 86.25 | |
| DGEDT+BERTBackbone=BERT2022.04 | 85.99 | |
| RGAT+BERTBackbone=BERT2022.04 | 85.91 | |
| InstructABSAvariant=22023.02 | 85.17 | |
| BERT+FinetuneBackbone=BERT2022.04 | 84.3 | |
| HGCNBackbone=GloVe + BiLSTM, Evaluation Protocol=Best performance2022.04 | 84.09 | |
| HGCNBackbone=GloVe + BiLSTM, Evaluation Protocol=Average performance2022.04 | 82.91 | |
| DualGCNBackbone=GloVe + BiLSTM2022.04 | 82.85 | |
| Dual-MRC2023.02 | 82.04 | |
| CDTBackbone=GloVe + BiLSTM2022.04 | 81.74 | |
| RGATBackbone=GloVe + BiLSTM2022.04 | 81.5 | |
| InterGCNBackbone=GloVe + BiLSTM2022.04 | 81.26 | |
| MCRF-SABackbone=GloVe + BiLSTM2022.04 | 80.94 | |
| DGEDTBackbone=GloVe + BiLSTM2022.04 | 80.89 | |
| BiGCNBackbone=GloVe + BiLSTM2022.04 | 80.83 | |
| ASGCNBackbone=GloVe + BiLSTM2022.04 | 80.8 | |
| AOABackbone=GloVe + BiLSTM2022.04 | 79.97 | |
| MGANBackbone=GloVe + BiLSTM2022.04 | 79.96 | |
| SA-LSTMBackbone=GloVe + BiLSTM2022.04 | 77.64 | |
| ATAE-LSTMBackbone=GloVe + BiLSTM2022.04 | 77.01 |