Aspect-level sentiment analysis on Rest15
85.61AccuracyHGCN+BERT
Evaluation results
| Method | Links | |
|---|---|---|
| HGCN+BERTBackbone=BERT, Evaluation Protocol=Best performance2022.04 | 85.61 | |
| InstructABSAvariant=22023.02 | 84.5 | |
| HGCN+BERTBackbone=BERT, Evaluation Protocol=Average performance2022.04 | 83.91 | |
| DualGCN+BERTBackbone=BERT2022.04 | 83.78 | |
| RGAT+BERTBackbone=BERT2022.04 | 83.15 | |
| InstructABSAvariant=12023.02 | 83.02 | |
| DGEDT+BERTBackbone=BERT2022.04 | 82.95 | |
| HGCNBackbone=GloVe + BiLSTM, Evaluation Protocol=Best performance2022.04 | 82.66 | |
| BERT+FinetuneBackbone=BERT2022.04 | 81.85 | |
| HGCNBackbone=GloVe + BiLSTM, Evaluation Protocol=Average performance2022.04 | 80.81 | |
| BiGCNBackbone=GloVe + BiLSTM2022.04 | 80.33 | |
| DualGCNBackbone=GloVe + BiLSTM2022.04 | 80.16 | |
| InterGCNBackbone=GloVe + BiLSTM2022.04 | 79.54 | |
| ASGCNBackbone=GloVe + BiLSTM2022.04 | 79.43 | |
| RGATBackbone=GloVe + BiLSTM2022.04 | 78.78 | |
| MCRF-SABackbone=GloVe + BiLSTM2022.04 | 78.71 | |
| DGEDTBackbone=GloVe + BiLSTM2022.04 | 78.32 | |
| AOABackbone=GloVe + BiLSTM2022.04 | 78.17 | |
| MGANBackbone=GloVe + BiLSTM2022.04 | 78.15 | |
| CDTBackbone=GloVe + BiLSTM2022.04 | 77.94 | |
| SA-LSTMBackbone=GloVe + BiLSTM2022.04 | 76.63 | |
| ATAE-LSTMBackbone=GloVe + BiLSTM2022.04 | 76.6 | |
| Dual-MRC2023.02 | 73.59 |