Table 9 Development (dev) set accuracy (%) of the machine learning models
SBERT before SMOTE | SBERT after SMOTE | |||
|---|---|---|---|---|
| Â | APC | ATM | APC | ATM |
Random forest | 65.9 ± 0.25 | 68.5 ± 0.68 | 51.4 ± 10.7 | 71.4 ± 1.16 |
XGBoost | 62.5 ± 0.29 | 73. ± 0.13 | 62.5 ± 0.29 | 73 ± 0.13 |
LightGBM | 64.9 ± 0.29 | 70.2 ± 0.64 | 64.9± 0.29 | 70.3 ± 0.64 |
CNN | 67.3 ± 0.04 | 71.1 ± 2.84 | 47.0 ± 17.4 | 69.4 ± 5.2 |
SimCSE before SMOTE | SimCSE after SMOTE | |||
|---|---|---|---|---|
| Â | APC | ATM | APC | ATM |
Random forest | 65.9 ± 0.15 | 73.2 ± 0.17 | 50.8 ± 10.9 | 71.6 ± 1.47 |
XGBoost | 62.5 ± 0.65 | 73.7 ± 0.17 | 62.5 ± 0.65 | 68.8 ± 0.73 |
LightGBM | 64.7 ± 0.29 | 74. ± 0.18 | 64.7 ± 0.29 | 70.7 ± 0.28 |
CNN | 67. ± 0.00 | 73 ± 0.02 | 43. ± 0.17 | 71 ± 0.04 |