Aspect-based Sentiment Analysis on SMM4H Task 2 2023 (test)
80Micro-F1DeBERTa-v3-base (augmented)
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DeBERTa-v3-base (augmented)Backbone=DeBERTa-v3-base, Data Augmentation=LLM-based augmented2026.03 | 80 | 65.9 | 46.7 | 87.4 | 63.6 | |
| RoBERTa-base (augmented)Backbone=RoBERTa-base, Data Augmentation=LLM-based augmented2026.03 | 78.5 | 63.9 | 46.1 | 86.9 | 58.7 | |
| RoBERTa-base (baseline)Backbone=RoBERTa-base, Data Augmentation=baseline2026.03 | 78.2 | 63.3 | 41.6 | 87.1 | 61.2 | |
| DeBERTa-v3-base (baseline)Backbone=DeBERTa-v3-base, Data Augmentation=baseline2026.03 | 78 | 65.6 | 46.6 | 86.9 | 63.3 | |
| BERT-base (augmented)Backbone=BERT-base, Data Augmentation=LLM-based augmented2026.03 | 76.8 | 61.9 | 42 | 86 | 57.7 | |
| BERT-base (baseline)Backbone=BERT-base, Data Augmentation=baseline2026.03 | 75.8 | 54.4 | 22.5 | 85.8 | 54.9 |