Efficient techniques for mining spatial databases (1206.0217v1)

Abstract: Clustering is one of the major tasks in data mining. In the last few years, Clustering of spatial data has received a lot of research attention. Spatial databases are components of many advanced information systems like geographic information systems VLSI design systems. In this thesis, we introduce several efficient algorithms for clustering spatial data. First, we present a grid-based clustering algorithm that has several advantages and comparable performance to the well known efficient clustering algorithm. The algorithm has several advantages. The algorithm does not require many input parameters. It requires only three parameters, the number of the points in the data space, the number of the cells in the grid and a percentage. The number of the cells in the grid reflects the accuracy that should be achieved by the algorithm. The algorithm is capable of discovering clusters of arbitrary shapes. The computational complexity of the algorithm is comparable to the complexity of the most efficient clustering algorithm. The algorithm has been implemented and tested against different ranges of database sizes. The performance results show that the running time of the algorithm is superior to the most well known algorithms (CLARANS [23]). The results show also that the performance of the algorithm do not degrade as the number of the data points increases.