摘要
为更新批量数据,提出一种基于DBSCAN的新聚类方法。该算法通过扫描原对象确定它们同增量对象间的关系,得到一个相关对象集,同时根据该相关对象和增量对象之间的关系获得新的聚类结果。实验结果表明,该算法与DBSCAN是等价的,能更有效地解决批量数据更新时的增量聚类问题。
In order to update the batch data, a novel clustering algorithm based on DBSCAN is proposed, which determines the relation between the original object and increment object by scanning the original one. Thus, a relevant object set is got, according to which the new clustering result is obtained combined with increment object. Experimental results show this algorithm is equal to DBSCAN, and can solve the increment clustering problem when the batch data is updated effectively.
出处
《计算机工程》
CAS
CSCD
北大核心
2009年第2期63-64,67,共3页
Computer Engineering
关键词
空间数据挖掘
增量聚类
空间数据库
批量更新聚类算法
spatial data mining
increment clustering
spatial database
Batch Update Clustering Algorithm（BUCA）
