[Retracted] Research on the Influence of DNN-Based Cross-Media Data Analysis on College Students’ New Media Literacy

Algorithm 1

Pseudocode of the optimization model.

	# Hyperparameters
	L = number of layers of 4 # neural network
	Learn_rate = 0.02 # Learning rate (in some books, learning rate is denoted by alpha)
	Iterators = 5000 # number of iterations
	n = Ea()
	N[0] = x.shape [0] # Layer 0 is the input layer.
	N[1] = 4 # Layer 1 has 4 nodes
	N[2] = 4 # Layer 2 has 4 nodes
	N[3] = 4 # Layer 3 has 4 nodes
	N[4] = 1 # Layer 4 1 node
	N[L] = 1 # force layer 1 node
	= Ea() # define the activation function hierarchically.
	[1] = linear
	[2] = linear
	[3] = linear
	[4] = linear
	D  = Ea(alpha) # define the activation function hierarchically.
	d [1] = dz_linear
	d [2] = dz_linear
	d [3] = dz_linear
	d [4] = dz_linear
	Loss = L2 # Define the cost function
	DA_loss = dAL2 # define cost derivatives