Table 3 Experimental results on the laptop dataset.
From: Graph-enhanced implicit aspect-level sentiment analysis based on multi-prompt fusion
| Â | Methods | Precision (%) | Recall (%) | F1 (%) |
|---|---|---|---|---|
Extractive methods | DP-ACOS | 13.0 | 5.70 | 8.0 |
Extract-Classify-ACOS | 45.56 | 29.48 | 35.80 | |
SGTS | 41.40 | 32.30 | 36.30 | |
Generative | Paraphrase | N/A | N/A | 43.51 |
Seq2Path | N/A | N/A | 42.97 | |
Special_Symbols+ UAUL | 44.38 | \(\underline{43.65}\) | \(\underline{44.01}\) | |
DLO + UAUL | 43.78 | 43.53 | 43.65 | |
ILO + UAUL | 43.92 | 43.46 | 43.69 | |
BART-CRN | 48.16 | 31.83 | 38.32 | |
The approach in this paper | \(\underline{46.12}\) | 44.84(\(\uparrow\) 1.19 ) | 45.47(\(\uparrow\) 1.37 ) |
