模糊HopField神经网络聚类及在DRP中的应用  

Clustering of Fuzzy Hopfield Neural Network and its Application in DRP

作　　者：陈维静[1] 张洪伟[1] 翁怀荣[1] 

机构地区：[1]四川大学,成都610065

出　　处：《计算机工程与应用》2006年第6期215-218,共4页Computer Engineering and Applications

摘　　要：探讨了模糊Hopfield神经网络的聚类功能及其阈值的灵敏度和上下确界。提出了一种基于该聚类理论的企业销售客户资信聚类模型,通过二分法实现自动搜索各个阈值确界的逼近值,并应用于一大型粮食业集团的DRP系统中的资信评估,采用ASP.NET+MSSQLServer2000的B/S构架实现,为该集团的销售客户资信等级分类提供了科学方法,其聚类结果的合理性得到了集团应用的验证。The fuzzy Hopfield neural network and its use of clustering are discussed.The sensitivity of threshold value and its supremum and infimum are researched.A credit clustering model based on FuzzyHN is proposed and implemented in the DRP system of a large farming corporation for Credit Rating.The supremum and infimum of threshold value are automatically searched by dichotomy in this model.It is designed with the use of asp.net plus MS SQL Server 2000 structure and provides a scientific method on credit clustering of the corporation＇s Suppliers.The rationality of results is verified by corporation＇s application.

关 键 词：模糊HOPFIELD神经网络 聚类 DRP 资信评估 二分法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]