Table 1. Performance comparison between different prediction methods and our proposed method.
| Feature extraction method |
Feature classification method | Accuracy (%) | Sensitivity (%) | Specificity (%) | AUC | False alarm rate (h−1) | Prediction rate (%) and time (min) | Training time (min) | References |
|---|---|---|---|---|---|---|---|---|---|
| MM-mDistEn | ANN | 98.66 | 91.82 | 99.11 | 0.84 | 0.014 | 99/26.73 | 3.5 | – |
| M-mDistEn | ANN | 88 | 85 | 90 | 0.8 | 0.081 | 90/23 | 3.7 | – |
| PE | ANN | 90 | 83 | 93 | 0.76 | 0.262 | 70/22 | 3.5 | – |
| N/A | MLP | 83.63 | 84.67 | 82.60 | N/A | 0.174 | N/A | 7.3 | (Daoud & Bayoumi, 2019) |
| DCNN | MLP | 95.41 | 92.8 | 94.1 | N/A | 0.072 | N/A | 12.5 | (Daoud & Bayoumi, 2019) |
| CNN | SVM | N/A | 92.7 | 90.8 | N/A | N/A | -/21 | N/A | (Usman, Khalid & Aslam, 2020) |
| N/A | CNN | 99.3 | N/A | 99.6 | N/A | 0.5 | N/A | N/A | (Gómez et al., 2020) |