BP(back propagation)神经网络是1986年由Rumelhart和McClelland为首的科学家提出的概念，是一种按照误差逆向传播算法训练的多层前馈神经网络，是目前应用最广泛的神经网络。(BP (back propagation) neural network was founded in 1986 by Rumelhart and McClelland, led by the scientists put forward the concept of error backward propagation algorithm is a kind of according to the training of multilayer feedforward neural network, is currently the most widely used neural network.)