Design and implementation for hopfield associative memory neural networks

• Main Authors: ZHANG Shaoping, XU Xiaozhong, MA Yan
• Format: Article
• Language: English
• Published: Academic Journals Center of Shanghai Normal University 2016-02-01
• Series: Journal of Shanghai Normal University (Natural Sciences)
• Subjects: hopfield neural network; associative memory capacity; numeral recognition; MATLAB; genetic Algorithm
• Online Access: http://qktg.shnu.edu.cn/zrb/shsfqkszrb/ch/reader/create_pdf.aspx?file_no=201601003&flag=1&year_id=2016&quarter_id=1

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.