Research Article
A BERT-Based Aspect-Level Sentiment Analysis Algorithm for Cross-Domain Text
Table 4
Results of Accuracy and F1 value.
| Source | Target | Accuracy | F1 value | NO | YES | NO | YES |
| B | D | 0.783 | 0.822 | 0.785 | 0.814 | E | 0.767 | 0.785 | 0.770 | 0.796 | K | 0.782 | 0.821 | 0.790 | 0.835 | V | 0.810 | 0.823 | 0.810 | 0.812 | B | 0.776 | 0.835 | 0.781 | 0.842 |
| D | E | 0.765 | 0.812 | 0.770 | 0.811 | K | 0.794 | 0.814 | 0.785 | 0.825 | V | 0.884 | 0.917 | 0.901 | 0.921 | B | 0.763 | 0.801 | 0.759 | 0.821 | D | 0.765 | 0.801 | 0.769 | 0.824 | K | 0.824 | 0.875 | 0.835 | 0.867 |
| E | V | 0.767 | 0.792 | 0,778 | 0.815 | B | 0.747 | 0.784 | 0.746 | 0.765 | D | 0.766 | 0.805 | 0.765 | 0.801 |
| K | E | 0.787 | 0.822 | 0.779 | 0.821 | V | 0.830 | 0.845 | 0.828 | 0.848 | B | 0.862 | 0.875 | 0.865 | 0.867 | D | 0.840 | 0.869 | 0.852 | 0.869 |
| V | E | 0.831 | 0.54 | 0.838 | 0.868 | K | 0.847 | 0.889 | 0.864 | 0.870 |
|
|