Table 9 Class based accuracy of the top-three classes in the YP dataset
From: Toward multi-label sentiment analysis: a transfer learning based approach
Class | Model | ||
|---|---|---|---|
XLNet | BERT | SVCÂ +Â CC | |
Food-Positive | 92.11 | 89.60 | 78.35 |
Food-Negative | 94.25 | 93.45 | 90.42 |
Misc-Positive | 97.06 | 97.25 | 89.12 |