Table 2 Hyper-parameters optimized with genetic algorithm
Name of hyper- parameters | Data types | Range | Interval |
|---|---|---|---|
Number of neurons in the input layer | Continuous | 4–100 | 1 |
Number of neurons in first hidden layer | Continuous | 4–100 | 1 |
Number of neurons in second hidden layer | Continuous | 4–100 | 1 |
Contain a dropout layer or not | Category | Yes/not | 1 |
Drop probability if there has a dropout | Floating point | 0–0.5 | 0.05 |
Epoch | Continuous | 30–1000 | 20 |
Distribution of regularization options | Category | 1–4 | 1 |
The number of features after PCA | Category | 5–31 | 1 |