Table 4.
The comparison of our study with sate-of-the-art algorithms. * means feature selection was performed before the classification.
| blackWork | Data Balancing | Channel | No. Data | Classifier | Acc | Se | Sp | AUC |
|---|---|---|---|---|---|---|---|---|
| black [27] | Yes (SMOTE) | CH3 | 262 term; 38 preterm | polynomial | – | 96% | 90% | 0.95 |
| [28] | Yes (Min–Max) | CH3 | 150 term; 19 preterm | SONIA | 92.7% | 91.2% | 94.5% | 0.93 |
| [29] | Yes (SMOTE) | CH3 | 262 term; 38 preterm | SVM | 87% | 96% | 79% | – |
| [30] | Yes (SMOTE) | CH3 | 262 term; 38 preterm | Combined * | – | 91% | 84% | 0.94 |
| [31] | Yes (ADASYN) | CH1-3 | 143 term; 19 preterm | RF * | 93% | 89% | 97% | 0.962 |
| [32] | Yes (ADASYN) | CH1 | 262 term; 38 preterm | SVM | 99.72% | 99.48% | 99.96% | – |
| [37] | Yes (SMOTE) | CH3 | 262 term; 38 preterm | Adaboost | – | – | – | 0.986 |
| [38] | No | CH1-2 | 26 term; 26 preterm | SVM * | 95.70% | 98.40% | 93% | 0.95 |
| [39] | Yes (ADASYN) | CH1-3 | 262 term; 38 preterm | SVM * | 96.25% | 95.08% | 97.33% | – |
| [33] | Yes (SMOTE) | CH1-3 | 275 term; 51 preterm | Ensemble * | 91.64% | 96.23% | 87.04% | 98.13 |
| [5] | Yes (SMOTE) | CH1-3 | 275 term; 51 preterm | LDA * | 89.2% | 98.4% | 79.9% | 0.936 |
| [34] | Yes (SMOTE) | CH1-3 | 262 term; 38 preterm | GBC | 85% | - | - | 0.91 |
| [35] | Yes (Partition-Synthesis) | CH1-3 | 275 term; 51 preterm | SVM * | 91% | 89.0% | 93% | 0.97 |
| [40] | Yes (ADASYN) | CH1-3 | 262 term; 38 preterm | SVM * | 98.5% | 98.4% | 98.4% | - |
| Ours | No | CH1 | 262 term; 38 preterm | SVM | 99.7% | 99.5% | 99.7% | 0.999 |