Atrial Fibrillation Prediction on PhysioNet 2017 (test)
94.88ROC AUCMINA
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MINAMechanism=Multilevel knowledge-guided attention2019.05 | 94.88 | 94.36 | 83.42 | — | — | |
| ExpertRFClassifier=Random Forest, Features=Expert features (rhythm, MAD, morphology, frequency)2019.05 | 93.94 | 88.16 | 81.8 | — | — | |
| ExpertLRClassifier=Logistic Regression, Features=Expert features (rhythm, MAD, morphology, frequency)2019.05 | 93.5 | 87.3 | 80.23 | — | — | |
| ACRNNAttention=Beat level and rhythm level2019.05 | 90.72 | 89.35 | 82.48 | — | — | |
| CRNNBackbone=CNN with bi-directional LSTM2019.05 | 90.4 | 89.43 | 82.62 | — | — | |
| CNNBackbone=CNN with global average pooling2019.05 | 87.11 | 86.69 | 79.14 | — | — | |
| MINA2020.12 | — | — | — | 7 | 46 | |
| TMFBackbone=VGG16-MLP2020.12 | — | — | — | 1 | 90 |