Table 4 The top-ten globally ranked features of the proposed model and LR
From: Interpretable clinical prediction via attention-based neural network
MLP_attention | LR | |||
|---|---|---|---|---|
Feature ID | Name | Frequency | Feature ID | Name |
21 | NT-proBNP | 736 | 21 | NT-proBNP |
94 | CCB | 374 | 7 | SBP |
78 | Left atrial diameter | 225 | 41 | Lactate dehydrogenase |
103 | Spironolactone | 180 | 57 | Monocytes ratio |
7 | SBP | 167 | 3 | Height |
32 | Sodium | 146 | 8 | DBP (Diastolic blood pressure) |
85 | Interventricular septal thickness | 146 | 5 | BMI |
87 | Left ventricular end systolic volume | 140 | 34 | Phosphorus |
10 | Anemia | 135 | 59 | Basophil ratio |
60 | Platelet count | 126 | 60 | Platelet count |