FRINGE:边界点的有效检测  

FRINGE:Efficient Detection Of Boundary Points

作　　者：陈阵[1] 于炯[1,2] 

机构地区：[1]新疆大学信息科学与工程学院,新疆乌鲁木齐830046 [2]北京理工大学计算机学院,北京100081

出　　处：《新疆大学学报（自然科学版）》2008年第3期263-268,共6页Journal of Xinjiang University(Natural Science Edition)

基　　金：国家自然科学基金(编号:60563002);教育部春晖计划项目(Z2005-1-65009);新疆自治区高校科研重点项目(XJEDU2004I03)

摘　　要：本文阐述了在数据集中找边界点的问题,边界点是分布在稠密数据集边缘的数据点.本文描述了一种称作FRINGE的新方法来检测边界点.FRINGE(an eFficient boundaRy poInts detectioN based on Grid and anglE)利用了网格技术和角度的特点,利用了具有多种特性的数据进行实验,实验结果表明FRINGE能在含有噪声点/孤立点的不同形状、大小的数据集上有效地检测出边界点,并且执行效率更高.This work addresses the problem of detecting boundary points in data sets. Boundary points are data points that distribute the edge of densely distributed data such as the cluster. It describes a novel approach called FRINGE （an eFficient boundary points detectioN based on Grid and anglE） to detect boundary points. FRINGE employs the grid technique and the angle feature. Experimental studies on data sets with varying characteristics indicate that FRINGE is able to detect boundary points in the noisy dataset containing different shapes and sizes clusters effectively and has higher efficiency.

关 键 词：边界点 边界格 密度 角度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]