Neural Network-Based Voting System with High Capacity and Low Computation for Intrusion Detection in SIEM/IDS Systems

Table 3

Parameters of the base neural network learners.

Parameter/base learner	NN1	NN2	NN3
No. of input features during training	0, 1, …, 29	30, 31, …, 59	50, 51, …, 89
No. of hidden layers	Single hidden layer
No. of hidden nodes	10	20	25
Optimizer	Stochastic gradient descent	Resilient back-propagation	Adam
Learning rate	1e − 3	3e − 4	1e − 4
Batch size	256	Full batch	512
Hidden activation function	RelU activation function
Loss function	Cross Entropy loss function
Activation function	Sigmoid activation function