论文部分内容阅读
作为人工神经网络的一种基本算法,BP网络在许多领域都有着广阔的应用前景。但由于其收敛速度太慢.因而很难投入实际应用。本文给出了几种改进的BP学习算法.如克服遗忘的BP(SBP)算法,变参数BP(PBP)算法,变换激励函数的BP(CBP)算法和改善梯度估计精度的BP(IBP)算法等,这些算法可以有效地提高其收敛速度。
As a basic algorithm of artificial neural network, BP network has broad application prospect in many fields. However, due to its slow convergence rate. It is difficult to put into practical application. In this paper, several improved BP learning algorithms are given. Such as overcoming the forgotten BP (SBP) algorithm, the variable parameter BP (PBP) algorithm, the BP (CBP) algorithm for transforming stimulus and the BP (IBP) algorithm for improving the accuracy of gradient estimation. These algorithms can effectively improve their convergence speed.