Research Article
DeeBERT-S3WD: Three-Way Multigranularity Decision for Interactive Information Sentiment Analysis Research
Table 2
Sentiment classification results.
| Method | Class | Total accuracy | Accuracy | F1 |
| SVM | Positive | 0.7694 | 0.7837 | 0.7549 | Negative | 0.7551 | 0.7396 |
| CNN | Positive | 0.7759 | 0.8063 | 0.7834 | Negative | 0.7455 | 0.7188 |
| DeeBERT | Positive | 0.8023 | 0.7953 | 0.7849 | Negative | 0.8092 | 0.8214 |
| DeeBERT-S3WD | Positive | 0.8177 | 0.8090 | 0.7951 | Negative | 0.8264 | 0.846 |
| BERT | Positive | 0.818 | 0.8102 | 0.8231 | Negative | 0.8257 | 0.8325 |
| RoBERTa | Positive | 0.8301 | 0.8215 | 0.8336 | Negative | 0.8387 | 0.8597 |
| DeBERTa | Positive | 0.8346 | 0.8324 | 0.8437 | Negative | 0.8367 | 0.8664 |
|
|