摘要
为了快速有效地检测聚类的边界点,提出了网格梯度、边界网格的概念以及一种基于网格梯度的边界点检测算法(Boundary Points Detecting Based Gradient of Grid,BPGG),该算法先求出网格的梯度值,根据此值判断该网格是否为边界网格,进而确定聚类边界点.实验表明该算法可以在含有任意形状、大规模数据集上快速有效地检测出聚类的边界点,并去除噪声.
In order to detect the boundary points of clusters, this paper proposes the concept of grid＇s gradient and boundary grid, and a boundary points detecting algorithm called （Boundary Points detecting based Gradien of Grid,BPGG）. BPGG detects boundary points of clusters according to grid＇s gradient. As shown by our experimental results, BPGG can be efficient and effective to detect boundary points and filter out noises in large datasets containing arbitary shaped clusters and noises.
出处
《微电子学与计算机》
CSCD
北大核心
2008年第3期77-80,共4页
Microelectronics & Computer
基金
国家自然科学基金项目(60673087)
郑州大学骨干教师基金项目
关键词
边界点
聚类
梯度
boundary points
dusters
gradient
