模糊神经网络分类器在企业建模知识管理中的应用  

Applying Fuzzy Neural Network Classifier to Knowledge Management in Enterprise Modeling

作　　者：沈晖[1] 陈禹六[1] 

机构地区：[1]清华大学自动化系,北京100084

出　　处：《系统工程理论与实践》2003年第11期1-7,118,共8页Systems Engineering-Theory & Practice

基　　金：国家 8 63计划 (2 0 0 1 AA41 1 1 3 0 ) ;欧盟 IST计划 (IST-2 0 0 0 -2 62 67)

摘　　要：为了对企业建模活动进行知识管理 ,开发了基于 Web的企业模型知识库系统 ,并按照 CIM-OSA的思想将参考模型库划分为通用、部分通用和专用三个层次 .如何在这三个层次之间建立合理的分类检索体系成为企业建模知识管理的一个核心问题 .文章引入模糊神经网络分类器解决企业参考模型库的模型分类问题 ,通过一个应用实例进行算法验证 ,并与基于模糊数据挖掘的分类方法进行比较分析 .实验数据表明 ,模糊神经网络分类器具有很好的性能 .In order to arrange knowledge management for enterprise modeling activities, a web-based enterprise models repository system was developed, in which, stemming from CIM-OSA′s three-level architecture, the organization of reference models was set as: Generic, Partial and Particular. How to establish a reasonable classification system among these three levels thus becomes a crucial problem for knowledge management in enterprise modeling. This paper introduced Fuzzy Neural Network Classifier to solve such a problem and provided a case study to illustrate the algorithms, comparing with another classification method based on fuzzy data mining. The experimental results showed that fuzzy neural network classifier has better performance.

关 键 词：企业建模 知识管理 模式发现 模糊聚类 模式识别 B-P神经网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]