Table 1 Comparative performance analysis of prior algorithms across multiple databases
From: mRMEBP: a unified framework for online detection of atrial fibrillation utilizing deep learning
Authors | Year | Main features | Database | Key techniques | Performances in % | |||
|---|---|---|---|---|---|---|---|---|
Se | Sp | PPV | ACC | |||||
Luo et al.34 | 2024 | RRIs | LTAFDB | Sobolev test statistics+End-to-end LM | 95.90 | 93.60 | − | 95.00 |
AFDB | 96.20 | 98.20 | − | 97.30 | ||||
MITDB | 93.80 | 95.60 | − | 95.40 | ||||
NSRDB | NA | 99.60 | NA | NA | ||||
Kumar et al.32 | 2023 | RRIs | AFDB | CNN | 96.06 | 98.29 | − | 97.04 |
MITDB | 96.87 | 86.94 | − | 87.98 | ||||
NSRDB | NA | 94.44 | NA | NA | ||||
Jahan et al.30 | 2022 | RRIs | LTAFDB | AdaBoost | 86.45 | 81.57 | − | − |
AFDB† | 87.58 | 89.27 | − | − | ||||
MITDB | 85.67 | 81.25 | − | − | ||||
NSRDB | NA | 93.21 | NA | NA | ||||
Liu et al.29 | 2022 | RRIs | AFDB | MGNN+Self-attention | 94.95 | 97.77 | 93.91 | 97.07 |
MITDB | 95.17 | 91.94 | 53.92 | 92.23 | ||||
NSRDB | NA | 96.86 | NA | NA | ||||
Andersen et al.23 | 2019 | RRIs | AFDB† | CNN+RNN+Median filter | 98.98 | 96.95 | 95.76 | 97.80 |
MITDB | 98.96 | 86.04 | 45.45 | 87.40 | ||||
NSRDB | NA | 95.01 | NA | NA | ||||
Sološenko et al.21 | 2019 | RRIs+PPIs | AFDB | SQA+Decision logic | 97.10 | 98.40 | − | 97.80 |
MITDB | 96.80 | 91.30 | − | 92.60 | ||||
NSRDB | NA | 99.20 | NA | NA | ||||
Ródenas et al.17 | 2017 | RRIs+AA | AFDB | COSEn+TQEn+LDA | 96.47 | 97.35 | − | 96.96 |
MITDB | 75.30 | 93.51 | − | 91.98 | ||||
Zhou et al.15 | 2015 | RRIs | LTAFDB⊤ | Symbolic dynamics+Shannon Entropy+Threshold | 96.14 | 95.73 | 97.03 | 95.97 |
AFDB | 97.37 | 98.44 | 97.89 | 97.99 | ||||
AFDB† | 97.31 | 98.28 | 97.89 | 97.84 | ||||
AFDB‡ | 98.43 | 98.46 | 97.92 | 98.45 | ||||
MITDB | 97.83 | 87.41 | 47.67 | 88.51 | ||||
NSRDB | NA | 99.68 | NA | NA | ||||
AFDB+NSRDB | 97.36 | 99.32 | 96.86 | 98.98 | ||||
AFDB†+NSRDB | 97.31 | 99.31 | 96.83 | 98.96 | ||||
AFDB‡+NSRDB | 98.43 | 99.35 | 96.82 | 99.19 | ||||
