论文部分内容阅读
本文提出了一种联想记忆网络权学习算法,文中将联想模型分解成一系列的非线性方程,针对这些非线性方程设计了一种能快速收敛的迭代方法,为了提高网络对缺损和噪声样本的联想记忆能力,通过提高网络阈值函数门限进行网络的严格训练,降低门限进行联想学习结果使得训练样本成为网络的稳定收敛点,并提高了网络的联想记忆能力。
This paper presents a learning algorithm for associative memory network. In this paper, the associative model is decomposed into a series of nonlinear equations. For the nonlinear equations, an iterative method that can converge rapidly is proposed. In order to improve the association between the neural network and the noise samples Memory ability, the threshold of the network function threshold to improve the network training, reduce the threshold for association learning results make the training samples become a stable convergence point of the network, and improve the associative memory ability of the network.