Title: EIDBSCAN: An Extended Improving DBSCAN algorithm with sampling techniques

Authors: Cheng-Fa Tsai, Chun-Yi Sung

Addresses: Department of Management Information Systems, National Pingtung University of Science and Technology, Pingtung 91201, Taiwan. ' Department of Management Information Systems, National Pingtung University of Science and Technology, Pingtung 91201, Taiwan

Abstract: Cluster analysis in data mining and knowledge discovery is an essential business application. This investigation describes a new clustering approach named EIDBSCAN that extends expansion seed selection into a sampling-based DBSCAN clustering algorithm. Additionally, the proposed algorithm may reduce eight Marked Boundary Objects to add expansion seeds according to far centrifugal force, which increases coverage. Our experimental results reveal that the proposed EIDBSCAN yields more accurate clustering results. In addition, in all the cases we studied, the proposed approach has a lower execution time cost than several existing well-known approaches, such as DBSCAN, IDBSCAN and KIDBSCAN clustering algorithms.

Keywords: data mining; data clustering; cluster analysis; algorithms; EIDBSCAN; extended improving DBSCAN; density-based spatial clustering of applications with noise; expansion seed selection; marked boundary objects; centrifugal forces; IDBSCAN; KIDBSCAN; business intelligence; knowledge discovery; sampling techniques.

DOI: 10.1504/IJBIDM.2010.030301

International Journal of Business Intelligence and Data Mining, 2010 Vol.5 No.1, pp.94 - 111

Available online: 14 Dec 2009 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article