Table 3 Optimum configuration parameter for the DNN model
From: Forest fire pattern and vulnerability mapping using deep learning in Nepal
Parameters | Values |
|---|---|
Number of nodes in the input layer | Number of input features |
Number of epochs | 106 |
Distribution | Gaussian |
Hidden layer | 150,150,150 |
Activation function | adam |
Input dropout ratio | 0.1 |
L1 | 1e−5 |
L2 | 0 |
Rho | 0.9 |