Table 3 Model parameter configuration for CNN-XGBoost fusion architecture.
From: Marine fishery resource dynamic prediction based on CNN-XGBoost fusion model
Parameter category | Parameter name | Parameter value |
|---|---|---|
CNN architecture | Convolutional Layers | 3 layers |
CNN filters | Filter Numbers | [32, 64, 128] |
CNN kernels | Kernel Sizes | [3, 5] |
CNN regularization | Dropout Rates | [0.3, 0.5] |
XGBoost trees | Maximum Depth | 6 |
XGBoost learning | Learning Rate | 0.1 |
XGBoost ensemble | Number of Estimators | 500 |
XGBoost regularization | L2 Regularization | 0.01 |
