Evaluation of the landslide susceptibility and its spatial difference in the whole Qinghai-Tibetan Plateau region by five learning algorithms

Geoscience Letters

Official Journal of the Asia Oceania Geosciences Society (AOGS)

Table 4 Summary of model structure and hyper-parameter used for building a deeply connected neural network

Hyperparameters	Setting values
Number of inputs parameters	13
Number of densely connected units	13
Activation function	ReLU
Layer dropout rate	0.4,0.3,0.2
Transfer function	Sigmoid
Loss function	Log Loss
Optimizer	SGD
Learning rate	0.001
Metrics	Accuracy
Epoch	300
Batch size	32
Validation split	0.2