期刊名称：International Journal of Computer Science & Technology
印刷版ISSN：2229-4333
电子版ISSN：0976-8491
出版年度：2011
卷号：2
期号：2(Version 1)
出版社：Ayushmaan Technologies
摘要：Spatial clustering techniques are a subset of clustering techniques applied on databases whose records have attributes intrinsically related to some spatial semantics. However, when applied to tasks like class identification on a spatial context, the traditional techniques might be unable to achieve good results, e.g. elements in the same cluster might not share enough similarity or the performance may be prohibitively poor. So far, a lot of spatial clustering algorithms have been proposed in many applications such as pattern recognition, data analysis, and image processing and so forth. However most of the well-known clustering algorithms have some drawbacks. To overcome these limitations, in this paper we propose a robust spatial clustering algorithm named NSCABDMDT (Novel Spatial Clustering Algorithm Based on distance matrices & Delaunay Triangulation). Delaunay diagram is used for determining neighbourhoods based on the neighbourhood notion, spatial association rules being defined. Whereas the distance matrices given the best method to determine distance between two data objects by increasing the efficiency of algorithm NSCABDMDT. Firstly, it even discovers arbitrary shape of cluster distribution. Secondly, like DBSCAN, Experiments show that NSCABDMDT does not require so much CPU processing time. Third it handles efficiently outliers.9?
