Introduction - If you have any usage issues, please Google them yourself
Neural network BP algorithm, learning rule is to use the steepest descent method, by back propagation to constantly adjust the network weights and threshold, minimize the error sum of squares of the network