一种使用启发式搜索优化的模糊规则分类方法  

A New Classification Method with Fuzzy Rules Based on Heuristics Search

作　　者：张燕[1] 袁书卿 鲁汉榕[3] 

机构地区：[1]空军雷达学院研究生队,武汉430019 [2]平顶山师范学院,河南平顶山467000 [3]空军雷达学院指挥自动化工程系,武汉430019

出　　处：《计算机工程与应用》2004年第1期99-102,共4页Computer Engineering and Applications

摘　　要：文章提出了一种基于模糊规则的分类方法。该方法首先介绍了基于模糊C均值聚类的模糊规则提取,然后利用所建立的模糊规则库设计了一种分类算法,并且利用启发式搜索来精简分类规则。使用IRIS数据对该文的方法进行了性能测试,结果表明该方法在训练样本较少的情况下,能得到很好的分类效果,并且通过规则精简,所使用的规则数目大大下降,而分类性能更加优良。A new classification method is proposed based on fuzzy rules in this p ap er.This method firstly introduces the fuzzy rules extraction technology based o n fuzzy C-means algorithm,then designs a new classification algorithm with the fuzzy rules,finally uses a heuristics algorithm to simplify the fuzzy rules.T he performance of the classification method is tested by the real IRIS data,and the results show that this algorithm has very good classification performance w ith fewer training samples,and the number of the fuzzy rules is decreased with the heuristics algorithm,meanwhile the classification performance is increased .

关 键 词：模糊规则 模糊C均值聚类 启发式搜索 

分 类 号：O159[理学—数学]