Research Article
Heartbeat Classification Using Normalized RR Intervals and Morphological Features
Table 3
Classification results for the imbalanced training dataset.
| Feature
configurations | N | S | V | Global | | Se1 | | Se2 | | Se3 | | Se | | Acc |
| | Nonnormalized RR intervals + WT features | 93.5 | 99.2 | 80.2 | 20.7 | 78.0 | 79.2 | 83.9 | 66.4 | 92.1 | | Normalized RR intervals + WT features | 94.8 | 99.1 | 77.2 | 24.5 | 78.2 | 78.3 | 83.4 | 67.3 | 93.2 | | Nonnormalized RR intervals + LP features | 92.7 | 99.1 | 84.1 | 20.6 | 78.2 | 76.4 | 85.0 | 65.4 | 91.4 | | Normalized RR intervals + LP features | 92.4 | 99.1 | 79.2 | 25.2 | 81.4 | 63.4 | 84.3 | 62.6 | 91.3 | | Nonnormalized RR intervals + WT + LP features | 93.7 | 99.0 | 82.9 | 20.6 | 75.6 | 84.7 | 84.1 | 68.1 | 92.2 | | Normalized RR intervals + WT + LP features | 94.6 | 98.9 | 78.2 | 24.1 | 77.8 | 79.7 | 83.5 | 67.6 | 93.0 |
|
|
WT: wavelet; LP: linear prediction.
|
