An Intrusion Detection Method Based on Fully Connected Recurrent Neural Network

Algorithm 1

Forward propagation algorithm.

	Input: the training sample was xi (i = l, 2, m), the weight matrix was Whx, Whh, and Wyh, the bias was bh and by, the activation function e uses the sigmoid function, and the classification function g uses the SoftMax function.
	Output: the output value corresponding to the training sample Xi
(1)	for xi from 1 to m do
(2)
(3)
(4)
(5)
(6)	End for