From: Aspect-level sentiment classification with fused local and global context
Methods | Restaurant | Laptop | MAMS | |||||
---|---|---|---|---|---|---|---|---|
Accuracy | Macro-F1 | Accuracy | Macro-F1 | Accuracy | Macro-F1 | Accuracy | Macro-F1 | |
PConvBERT | 86.96 | 80.87 | 81.66 | 78.33 | 76.73 | 75.82 | 84.36 | 83.95 |
PConvBERT Â w/o LSFE | 85.36 | 78.70 | 79.15 | 74.87 | 74.86 | 74.08 | 83.08 | 82.52 |
PConvBERT Â w/o Focal Loss | 85.36 | 78.15 | 79.31 | 75.34 | 74.46 | 72.83 | 83.76 | 83.26 |
PConvBERT w/o FGM | 86.34 | 79.35 | 80.56 | 77.83 | 75.00 | 73.92 | 82.93 | 82.48 |
PConvBERT Â w/o Prompt | 85.36 | 79.13 | 76.96 | 70.74 | 75.43 | 74.24 | 83.08 | 82.53 |
P ConvRoBERTa | 89.29 | 84.27 | 83.54 | 80.89 | 78.47 | 77.53 | 85.55 | 85.05 |
PConvRoBERTa  w/o LSFE | 87.86 | 80.65 | 82.29 | 79.31 | 76.01 | 75.00 | 83.98 | 83.57 |
P ConvRoBERTa  w/o Focal Loss | 85.62 | 77.97 | 82.45 | 79.72 | 76.30 | 75.28 | 83.23 | 82.62 |
PConvRoBERTa  w/o FGM | 87.77 | 81.56 | 82.92 | 80.11 | 75.87 | 75.04 | 83.01 | 82.44 |
PConvRoBERTa  w/o Prompt | 86.43 | 78.70 | 83.07 | 80.13 | 73.84 | 72.84 | 84.88 | 84.35 |