WSN-DS: A Dataset for Intrusion Detection Systems in Wireless Sensor Networks
Table 7
Parameters for MLP neural network classifier.
Parameter
Explanation
Used value
L
Learning rate: used for weight adjustment on each iteration. (The value should be between 0 and 1.)
0.3
M
Momentum: used for weight adjustment during backpropagation, in order to speed up convergence and avoid local minima. (The value should be between 0 and 1.)
0.2
N
The number of epochs or passes through training data.
500
V
The percentage of the validation set from the training data.
20%
S
Seed for random number generator. Random numbers are used for setting initial weights for the connections between nodes. (The value should be ≥0.)
0
E
Threshold for consecutive errors allowed during validation testing before the neural network terminates. (The value should be >0.)
20
H
Number of nodes in the hidden layer which is represented as follows: number of hidden layers (number of neurons in each layer).
