Research Article
Combining Rhythm Information between Heartbeats and BiLSTM-Treg Algorithm for Intelligent Beat Classification of Arrhythmia
Table 4
Comparison of the classification results of RNN, GRU, and LSTM network models.
| ā | RNN | GRU | LSTM | Se (%) | Sp (%) | +p (%) | Acc (%) | Se (%) | Sp (%) | +p (%) | Acc (%) | Se (%) | Sp (%) | +p (%) | Acc (%) |
| N | 99.61 | 97.10 | 99.35 | 99.15 | 99.45 | 97.80 | 99.51 | 99.15 | 99.69 | 97.40 | 99.42 | 99.27 | S | 90.40 | 99.80 | 92.86 | 99.54 | 92.72 | 99.77 | 92.11 | 99.58 | 92.05 | 99.90 | 96.19 | 99.68 | V | 97.66 | 99.80 | 97.26 | 99.66 | 97.94 | 99.76 | 96.61 | 99.63 | 96.98 | 99.77 | 96.84 | 99.59 | F | 71.26 | 99.94 | 91.18 | 99.72 | 71.26 | 99.91 | 86.11 | 99.68 | 75.86 | 99.91 | 86.84 | 99.72 | Q | 99.43 | 99.94 | 99.32 | 99.90 | 99.77 | 99.90 | 98.88 | 99.89 | 99.55 | 99.97 | 99.66 | 99.94 |
|
|