H-SCAN : A Hash-based Spatial Clustering Algorithm for Knowledge Extraction
O Byeong-U, Gi-Jun Han · Jeongbo gwahaghoe nonmunji. so'peuteuweeo mich eung'yong · 1999
Recently, the necessity of spatial data mining which is the extraction of implicit knowledge, spatial relationships, or other knowledge not explicitly stored in spatial databases has been increased due to a huge amount of spatial data. Spatial clustering algorithms have been proposed in recent years for efficient spatial data mining. In this paper, we propose an efficient spatial clustering algorithm, named H-SCAN (A Hash-based Spatial Clustering Algorithm for kNowledge Extraction), which overcomes the disadvantages of existing spatial clustering algorithms. H-SCAN can perform efficient spatial clustering of large spatial data for practical spatial data mining by storing spatial data in an object-oriented database, handling noise problem, and processing line and polygon spatial objects as well as point spatial objects. In addition, it can find arbitrary shape of clusters such as a cluster that has holes inside or nested clusters for complex distribution of spatial data, integrates a spatial clustering structure and a spatial index structure, and utilizes read-only property of spatial data mining for the efficient knowledge extraction.