一种基于Tabu搜索的模糊聚类方法  

Fuzzy clustering method based on Tabu search

作　　者：刘素华[1] 

机构地区：[1]河南工业大学计算机科学系,河南郑州450052

出　　处：《计算机应用》2005年第1期42-44,48,共4页journal of Computer Applications

基　　金：河南省科技攻关项目(0224010011)

摘　　要：首先,对模糊C 均值聚类算法做了简要分析和评论,在此基础上,将Tabu搜索引入模糊聚类,以克服模糊C 均值聚类的局部性和对初始聚类中心的敏感性,采用了适合于模糊聚类的树型编码方案。然后,给出了新算法的实现方法及步骤。仿真实验表明,新方法在速度和解的质量方面都达到了令人满意的效果。The fuzzy C-means clustering algorithm was described and remarked. On the basis of systematic analysis of current algorithms, tabu search was introduced into fuzzy clustering to solve the locality and the sensitiveness of the initial condition of fuzzy C-means clustering. A tree encoding scheme proper for the fuzzy clustering algorithm was determined. The design method and steps of the new algorithm were provided. Simulations results show that this method performs well in both the quality of obtained feature subset and efficiency.

关 键 词：模糊C-均值聚类 TABU搜索 树型编码 模式识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]