Research Article
Combining Rhythm Information between Heartbeats and BiLSTM-Treg Algorithm for Intelligent Beat Classification of Arrhythmia
Table 9
Comparison with other studies.
| | | Reference | Classifier | Performance (%) |
| | Feature engineering | Yang et al., 2021 [9] | Random forest | Acc = 98.1
Se = 75.2
+p = 93.9 | | Ji et al., 2021 [15] | Stacking-DWKNN | Acc = 99.01
Sen = 99.65; Spn = 94.94; +Pn = 99.38
Ses = 89.42; Sps = 99.85; +Ps = 94.90
Sev = 97.21; Spv = 99.78; +Pv = 97.07
Sef = 80.77; Spf = 99.94; +Pf = 88.73 | | Zhu et al., 2018 [7] | SVM | Acc = 97.80
Sen = 99.27; +Pn = 98.48
Ses = 87.47; +Ps = 95.25
Sev = 94.71; +Pv = 95.22
Sef = 73.88; +Pf = 86.09 | | Deep learning | Pandey et al., 2017 [21] | 9-layer CNN | Acc = 94.03
Sen = 91.54; Spn = 96.71; +Pn = 87.43
Ses = 90.59; Sps = 98.63; +Ps = 94.30
Sev = 94.22; Spv = 98.84; +Pv = 95.30
Sef = 96.06; Spf = 98.67; +Pf = 94.76
Seq = 97.75; Spq = 99.69; +Pq = 98.73 | | Ji et al., 2019 [20] | 1D-CNN | Acc = 99.21
Sen = 98.27; Spn = 99.39
Sev = 97.54; Spv = 99.44
Sef = 98.07; Spf = 99.50 | | Wu et al., 2020 [23] | CNN-BiLSTM | Acc = 97.29
Sen = 98.57; Spn = 93.62; +Pn = 98.81
Ses = 84.97; Sps = 99.13; +Ps = 82.80
Sev = 94.90; Spv = 99.35; +Pv = 94.00
Sef = 76.89; Spf = 99.77; +Pf = 80.45 | | Pandey et al., 2021 [25] | BiLSTM | Acc = 98.58
Sen = 99.54; +Pn = 99.44
Ses = 92.00; +Ps = 91.02
Sev = 95.81; +Pv = 96.80
Sef = 80.55; +Pf = 85.22 | | Proposed | BiLSTM-Treg | Acc = 99.32
Sen = 99.78; Spn = 97.95; +Pn = 99.54
Ses = 93.38; Sps = 99.91; +Ps = 96.58
Sev = 98.63; Spv = 99.83; +Pv = 97.69
Sef = 72.41; Spf = 99.96; +Pf = 94.03
Seq = 100.00; Spq = 99.98; +Pq = 99.77 |
|
|
Bold values represent the best experimental results which correspond to the evaluation criteria for one certain type.
|
