Table 4 Performance of the classification of randomly selected testing heartbeat subsets across the entirety of both MIT-BIH and INCART databases.
From: ECG heartbeat classification using progressive moving average transform
Experiment | Database | Class | Precision | Sensitivity | Specificity | F1-score | AUC | Accuracy |
|---|---|---|---|---|---|---|---|---|
Exp.1 | MIT-BIH | N | 99.43% | 99.65% | 92.81% | 99.54% | 1.00 | 99.09% |
S | 87.50% | 76.09% | 99.86% | 81.40% | 0.99 | |||
V | 95.83% | 95.08% | 99.73% | 95.45% | 1.00 | |||
Macro average | 94.25% | 90.27% | 97.47% | 92.13% | 0.99 | |||
Exp. 2 | INCART | N | 98.85% | 99.58% | 91.62% | 99.22% | 1.00 | 98.37% |
S | 50.29% | 39.91% | 98.81% | 44.50% | 1.00 | |||
V | 96.63% | 92.23% | 96.20% | 94.39% | 1.00 | |||
Macro average | 81.72% | 77.24% | 95.54% | 79.37% | 1.00 |
