【摘要】 本文针对学习CP网络在训练过程中出现的不稳定性提出了一种改进的CP网络学习算法。由于这种改进的学习算法在训练过程中将希望输出的信息传递给第一层权值，因此它有效地克服了CP网络训练过程中的不稳定性，避免了网络在训练过程中陷于局部最小点。本文最后给出了一个实例，通过这一实例可以看出改进的学习算法明显地优于传统的学习算法。更多还原

【Abstract】 Aimed at the instability of CP network in the training process, an improved learning algorithm for CP network is proposed in this paper. Due to expected output information being propagated to the first layer of weights in the improved learning algorithm,the instability is effectively overcomed, so it avoids the local minimum appearing in the training process of CP network. At the end of this paper an example is provided. The result shows that the improved learning algorithm is obviously superior to the traditional learning algorithm.更多还原