This algorithm can improve the astringency and increment learning ability of the neural network effectively.

• 利用该算法对神经网络的权值和域值进行训练,有效地提高神经网络的收敛性及增量学习能力。