Research Article
A BERT-Based Aspect-Level Sentiment Analysis Algorithm for Cross-Domain Text
Table 5
Accuracy of different methods.
| Source | Target | Method | SCL-ML | ITIAD | CGRU | Ours |
| B | D | 0.734 | 0.798 | 0.815 | 0.822 | E | 0.723 | 0.782 | 0.778 | 0.785 | K | 0.762 | 0.805 | 0.797 | 0.821 | V | 0.752 | 0.764 | 0.785 | 0.823 | B | 0.792 | 0.814 | 0.825 | 0.835 |
| D | E | 0.771 | 0.787 | 0.805 | 0.812 | K | 0.765 | 0.786 | 0.796 | 0.814 | V | 0.806 | 0.835 | 0.855 | 0.917 | B | 0.784 | 0.794 | 0.797 | 0.801 | D | 0.798 | 0.735 | 0.781 | 0.801 |
| E | K | 0.704 | 0.787 | 0.826 | 0.875 | V | 0.708 | 0.756 | 0.775 | 0.792 | B | 0.762 | 0.764 | 0.761 | 0.784 | D | 0.793 | 0.775 | 0.802 | 0.805 |
| K | E | 0.724 | 0.769 | 0.809 | 0.822 | V | 0.751 | 0.755 | 0.817 | 0.845 | B | 0.824 | 0.765 | 0.842 | 0.875 | D | 0.817 | 0.842 | 0.868 | 0.869 |
| V | E | 0.778 | 0.836 | 0.832 | 0.854 | K | 0.813 | 0.854 | 0.867 | 0.889 |
|
|