Aspect Based Sentiment Analysis Absa On Acos
평가 지표
F1 (Laptop)
F1 (Restaurant)
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | F1 (Laptop) | F1 (Restaurant) | Paper Title | Repository |
|---|---|---|---|---|
| Extract-Classify | 36.42 | 43.77 | Aspect-Category-Opinion-Sentiment Quadruple Extraction with Implicit Aspects and Opinions | |
| ChatGPT (gpt-3.5-turbo, few-shot) | - | 37.71 | MvP: Multi-view Prompting Improves Aspect Sentiment Tuple Prediction | |
| Paraphrase | 43.51 | 61.16 | Aspect Sentiment Quad Prediction as Paraphrase Generation | |
| MvP (muilti-task) | 43.84 | 60.36 | MvP: Multi-view Prompting Improves Aspect Sentiment Tuple Prediction | |
| DLO | 43.64 | 59.99 | Improving Aspect Sentiment Quad Prediction via Template-Order Data Augmentation | |
| ChatGPT (gpt-3.5-turbo, zero-shot) | - | 27.11 | MvP: Multi-view Prompting Improves Aspect Sentiment Tuple Prediction | |
| TAS-BERT | 27.31 | 33.53 | Aspect Sentiment Quad Prediction as Paraphrase Generation | |
| MvP | 43.92 | 61.54 | MvP: Multi-view Prompting Improves Aspect Sentiment Tuple Prediction | |
| UnifiedABSA (multi-task) | 42.58 | 60.60 | UnifiedABSA: A Unified ABSA Framework Based on Multi-task Instruction Tuning | - |
0 of 9 row(s) selected.