基于聚类有效性分析的模糊粗糙集归纳学习方法  

Inductive Learning Approach of Fuzzy-rough Set Based on Clustering Valid Analysis

作　　者：印勇[1] 孙如英[1] 

机构地区：[1]重庆大学通信工程学院,重庆400044

出　　处：《计算机工程》2008年第10期86-88,共3页Computer Engineering

基　　金：重庆市应用基础研究基金资助项目(6976)

摘　　要：引入模糊C均值聚类算法进行连续属性模糊化,通过聚类有效性分析来确定最佳分类数目,克服了属性模糊化方法需要人为确定划分类数的缺点。用属性模糊化得到的属性隶属度矩阵约简模糊粗糙属性,由此提出一种基于模糊粗糙集的属性约简算法。实例验证了该方法的可行性和有效性。Fuzzy C means clustering is introduced to fuzzify the continuous attribute, and the best minute class number is obtained by the valid analysis of clustering. It has overcome the disadvantage of determining artificially the class number for fuzzifing attribute approach. The attribute degree of membership matrix which obtained by attribute fuzzified is used to attributes reduction, and attributes reduction algorithm based on fuzzy rough sets is given. An example is illustrated to prove its feasibility and effectiveness.

关 键 词：模糊粗糙集 模糊聚类 有效性分析 决策表 属性约简 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]