Aspect Based Sentiment Analysis Absa On Tasd
평가 지표
F1 (R15)
F1 (R16)
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | F1 (R15) | F1 (R16) | Paper Title | Repository |
|---|---|---|---|---|
| LEGO-ABSA (multi-task) | 62.30 | 71.80 | LEGO-ABSA: A Prompt-based Task Assemblable Unified Generative Framework for Multi-task Aspect-based Sentiment Analysis | - |
| MvP (multi-task) | 64.74 | 70.18 | MvP: Multi-view Prompting Improves Aspect Sentiment Tuple Prediction | |
| ChatGPT (gpt-3.5-turbo, zero-shot) | - | 34.08 | MvP: Multi-view Prompting Improves Aspect Sentiment Tuple Prediction | |
| Paraphrase | 63.06 | 71.97 | Aspect Sentiment Quad Prediction as Paraphrase Generation | |
| MvP | 64.53 | 72.76 | MvP: Multi-view Prompting Improves Aspect Sentiment Tuple Prediction | |
| TAS-BERT | 57.51 | 65.89 | Aspect Sentiment Quad Prediction as Paraphrase Generation | |
| GAS | 60.63 | 68.31 | Towards Generative Aspect-Based Sentiment Analysis | |
| ChatGPT (gpt-3.5-turbo, few-shot) | - | 46.51 | MvP: Multi-view Prompting Improves Aspect Sentiment Tuple Prediction | |
| DLO | 62.95 | 71.79 | Improving Aspect Sentiment Quad Prediction via Template-Order Data Augmentation |
0 of 9 row(s) selected.