A Novel Hybrid Spatial Clustering Algorithm

Clustering spatial data is a well-known problem that has been extensively studied to find hidden patterns or meaningful sub-groups, it is the problem of grouping data based on similarity and has many applications such as satellite imagery, geographic information systems, medical image analysis, pattern recognition, data clustering and signal processing. While this problem has attracted the attention of many researchers for many years. In this paper we discuss some very recent clustering approaches and present our research on spatial clustering approaches. Many algorithms have been designed to detect clusters in spatial databases. In this algorithm, clusters are detected and adjusted according to the intra-relationship within clusters and the interrelationship between clusters and outliers, and vice versa. The experimental results prove that CLARA clustering algorithm when combined with medoid improves the accuracy of detection and increases the time efficiency.