点击次数：

论文类型：期刊论文

发表时间：2008-12-01

发表刊物：Journal of Computational Information Systems

收录刊物：EI、Scopus

卷号：4

期号：6

页面范围：2473-2481

ISSN号：15539105

摘要：Thanks to their capabilities to handle clusters with arbitrary shape, Density-based clustering algorithms play important roles in various fields such as pattern recognition and so on. However, most Density-based clustering algorithms are sensitive to the parameters, and have difficulties in dealing with various density datasets which appear frequently in reality. In this paper, a novel Density-based clustering algorithm, Density-Tag Based Clustering (DTBC), is proposed. DTBC uses density-tag to automatically detect density distribution of the dataset, thus is effective for various density dataset. Meanwhile, DTBC needs only one parameter to which it is not sensitive. Experimental results show that DTBC is more effective for various density dataset than other typical Density-based clustering algorithms such as DBSCAN and KNNCLUST. © 2008 Binary Information Press.