“…BP neural network is a method which has teachers, input P learning samples: x1,x2, ,xp, and its related teachers are: T1,T2, ,Tp. Learning algorithm is to use the error of actual output of network Y1,Y2, ,Yp and teachers T1,T2, Tp to modify the connection of network and threshold, and it should make YP1 be close to de requiring Tp1 in every possible way [10,11].…”