Table 2 Details and parameters of the selected ANN
From: A neural network-based framework for financial model calibration
Parameters | Options |
|---|---|
Hidden layers | 4 |
Neurons(each layer) | 200 |
Activation | ReLu |
Dropout rate | 0.0 |
Batch-normalization | No |
Initialization | Glorot_uniform |
Optimizer | Adam |
Batch size | 1024 |