Detection of clusters in Spatial Databases is a major task for knowledge discovery. Density based clustering algorithms plays a major role in this domain. DBSCAN algorithm effectively manages to detect clusters of any arbitrary shape with noise, but it fails to detect local clusters. DDSC and LDBSCAN does manages to detect local clusters effectively, but the number of input parameters are high. In this paper we have proposed a new density based clustering algorithm which introduces a concept called Cluster Constant. It basically represent the uniformity of distribution of points in a cluster. The proposed algorithm has minimized the input to be provided by the user down to one parameter (Minpts) and has made the other parameter (Eps) adaptive. Further we have also used some heuristics in order to improve the running time of the algorithm. Experiment results shows that the proposed algorithm detects local clusters of any arbitrary shape very effectively and also improves the running time of the algorithm. © 2011 Springer-Verlag.
CITATION STYLE
Tripathy, A., Maji, S. K., & Patra, P. K. (2011). UDSCA: Uniform Distribution based Spatial Clustering Algorithm. In Communications in Computer and Information Science (Vol. 190 CCIS, pp. 649–660). https://doi.org/10.1007/978-3-642-22709-7_63
Mendeley helps you to discover research relevant for your work.