基于联想记忆的Hopfield神经网络的设计与实现

介绍了反馈型神经网络Hopfield网络的定义、原理、模型和基本学习规则,并构造了一个用于联想记忆的Hopfield神经网络模型.对实验结果进行分析、比较,实验结果表明:Hopfield神经网络用于数字识别是可行、有效的;该方法较传统神经网络能提高网络的记忆能力和数字识别的正确率;该方法有别于以往的BP神经网络的模式识别,结合一些优化算法,如遗传算法,能对Hopfield神经网络的联想记忆稳态进行优化,增强神经网络的联想记忆能力.

Design and implementation for hopfield associative memory neural networks

This paper introduces the definition,principle,model and basic learning rules of feedback neural network,i.e.Hopfield network,and constructs a Hopfield neural network model for associative memory.The experimental results are analyzed and compared.The results show that Hopfield neural network for digital identification is feasible and effective.This method,compared with that of traditional neural network,can improve the network memory capacity and accuracy of digital identification.This method is different from the previous BP neural network pattern recognition,and can optimize associative memory steady-state of Hopfield neural network and enhance the capacity of the associative memory of neural network in combination with some optimization algorithms such as genetic algorithm.