Table 10 tenfold cross-validation of the proposed ensemble model based on stacking model contrasted with other models.
Model | Accuracy-mean | Standard-deviation | Macro-precision-mean | Standard-deviation | Macro-recall-mean | Standard-deviation | F1score-mean | Standard-deviation |
|---|---|---|---|---|---|---|---|---|
SVM | 90 | 0.02 | 90 | 0.04 | 91 | 0.04 | 91 | 0.04 |
LR | 86% | 0.04 | 88% | 0.05 | 90% | 0.04 | 89% | 0.05 |
DT | 84% | 0.03 | 86% | 0.04 | 88% | 0.03 | 87% | 0.04 |
GB | 88% | 0.04 | 89% | 0.06 | 89% | 0.04 | 88% | 0.04 |
XSTM | 92% | 0.03 | 88% | 0.03 | 91% | 0.02 | 90% | 0.04 |
Mamba | 95% | 0.02 | 89% | 0.06 | 89% | 0.03 | 88% | 0.06 |
BERT (transformer) | 89% | 0.04 | 90% | 0.04 | 90% | 0.04 | 89% | 0.03 |
Proposed (stacking-based ensemble method) | 96% | 0.02 | 93% | 0.02 | 91% | 0.03 | 92% | 0.02 |
