Algorithm 2: Backpropagation algorithm.

(1)The value of weights in all layers are initialized randomly. (2)The values of each neuron in the hidden layer and output layer are calculated. (3)The weights in a neural network are updated using Levenberg–Marquardt optimization. (4) Step 2 is repeated until one of the following conditions is achieved: Reaching the maximum number of epochs. Exceeding the maximum specified time. Achieving the target performance.