论文部分内容阅读
该文将局域互联神经网络的概念推广到三阶的情况，并对三阶局域互联神经网络的关联存储特性进行了计算机模拟。结果表明，与传统的三阶全局互联神经网络相比，三阶局域网的互联权重矩阵大大减小了，同时基本保持原有的全局关联存储能力，故更具实际应用价值。
This paper generalizes the concept of local interconnection neural network to the third order, and simulates the related storage characteristics of the third order local interconnection neural network. The results show that compared with the traditional third-order global interconnection neural network, the interconnection weight matrix of the third-order local area network is greatly reduced, while the original global associative storage capability is basically maintained. Therefore, it is more practical and practical.