Table 9 The selected features of deep features
From: An improved electrocardiogram arrhythmia classification performance with feature optimization
Model | Classification | Cases | Number of extracted features | Number of selected features |
|---|---|---|---|---|
6 | NSR and AF | AE + TLO + single layer perceptron | 5000 | 4 |
7 | NSR and AF | CNN + TLO + fully connected layer | 5000 × 8 | 277 |
8 | NSR, AF, SBR, and SAR | AE + TLO + single layer perceptron | 5000 | 4 |
9 | NSR, AF, SBR, and SAR | CNN + TLO + fully connected layer | 5000 × 8 | 91 |