一种高效的基于联合熵的边界点检测算法  

An efficient boundary points detecting algorithm based on joint entropy

作　　者：邱保志[1] 曹鹤玲[1] 

机构地区：[1]郑州大学信息工程学院,郑州450001

出　　处：《控制与决策》2011年第1期71-74,79,共5页Control and Decision

基　　金：国家自然科学基金项(60673087);河南省教育厅自然科学基金项目(2009A520028)

摘　　要：为了快速有效地检测出聚类的边界点,提出一种将网格技术与联合熵相结合的边界点检测算法.该算法中网格技术用于快速查找数据集中聚类边界所在的网格范围,联合熵用于在边界落入的网格范围内准确识别聚类的边界点.实验结果表明,该算法能够在含有噪声点/孤立点的数据集上,有效地检测出聚类的边界,运行效率高.In order to detect boundary points of clusters quickly and efficiently, a boundary points detecting algorithm（EDGE） is proposed, which employs grid technique and joint entropy. Grid technique is used to search the scope of grids which the boundary of clusters is located in and joint entropy is used to detect boundary points of clusters in these grids. The experimental results show that EDGE can detect boundary points of clusters in datasets with noises/outliers effectively and efficiently,

关 键 词：边界点 联合熵 网格 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]