Table 4 Impact of different learning rates and activation function ReLU on the performance of the DNN model using a fivefold model
From: N6-methyladenine identification using deep learning and discriminative feature integration
Species | LR | ReLU ACC (%) | Tanh ACC (%) |
|---|---|---|---|
F. vesca | 0.01 | 97.70 | 94.91 |
0.02 | 97.34 | 94.34 | |
0.03 | 96.87 | 93.67 | |
R. chinensis | 0.01 | 95.75 | 92.65 |
0.02 | 95.05 | 91.17 | |
0.03 | 95.53 | 90.94 |