Modeling the Spread of COVID-19 Infection Using a Multilayer Perceptron

Table 2

Hyperparameters of MLPs for best models achieved. Each column is one of the models—predicting the number of infected, recovered, and deceased patients. Hyperparameters that resulted in the best model are shown in rows.

Hyperparameter	Infected model	Recovered model	Deceased model
Solver	LBFGS	LBFGS	LBFGS
Initial learning rate	0.1	0.5	0.01
Learning rate adjustment	Constant	Constant	Adaptive
Hidden layer tuple	(4, 4, 4, 4)	(4, 4, 4, 4)	(4, 4, 4, 4)
Activation function	ReLU	ReLU	ReLU
L2 regularization parameter	0.0001	0.001	0.01