A Spatial Polygon Objects Clustering Algorithm Based on Topological Relations for GML Data

Spatial clustering is an important research topic in spatial data mining. This paper proposes algorithm SPCTR-GML for clustering spatial polygon objects based on topological relations for GML data. In the algorithm, we firstly use the idea of grid to compute all topological relations including contains, intersection and adjacent relations among spatial objects, and then a novel method is defined to measure the similarity between spatial polygon objects based on the topological relations, finally polygon objects are clustered using algorithm ROCK. Experimental results show that SPCTR-GML is effective and efficient.