Research on the Interaction of Genetic Algorithm in Assisted Composition

Algorithm 2

The pseudocode for the network model training algorithm.

	Input: training data
	Learning rate:
	The connection weights and thresholds of neurons are initialized in (0, 1).
	REPEAT
	FOR ALL IN
	Step1: Calculate the data of the current sample according to
	Step2: Calculate the gradient term of neurons in the output layer according to
	Step3: Compute the gradient term of hidden layer neuron according to
	Step 4: Update connection weights and threshold according to
	END
	UNTIL Stop condition met
	OUTPUT = Network model