“…where Wij is weight between the hidden layers and input layers, i and i are the number of neuron of hidden layer and input layer respectively, 8 i is the bias of hidden layer 1548 neurons; P is the number of hidden layer neurons, in order to simulate the characteristics of the nonlinear of biological neurons, U i is used as independent variables of S-type function (Sigmoid function), Yi is the output of the hidden layer neurons, S-function is as following: 4) The desired output dt and the actual output of neural network Y t is compared, the error is calculated [9], mean square error is used as following m e=1/2�:cdt-yt)2 (5) if e is less than q (q i&=predetermined value), the algorithm is end, otherwise go to the steps (5), that is back-propagation.…”