Abstract: Clustering is an important data mining method for knowledge discovery in large databases. It is an exploratory data analysis tool which aims at categorizing different objects into groups (clusters) in such a way that the degree of association between two objects is high if they belong to the same group and low otherwise. Discovering clusters in spatial data is a challenging one because of its complexity nature. The clusters in spatial data are of different sizes, shapes and densities, and also contain noise and outliers. Different clustering techniques have been proposed for knowledge discovery from spatial databases. DBSCAN (Density Based Spatial Clustering of Applications with Noise) algorithm is a traditional and well known density-based clustering method. It defines a cluster as a maximal set of density-connected points. It can detect clusters of arbitrary shapes and filter out noise effectively. The clusters which are formed based on density are easy understandable and it does not limit the shapes. Besides its popularity, DBSCAN needs some improvements for better results. In this paper, we have analyzed and presented various significant enhancements of DBSCAN algorithm for our evaluation.

Keywords: Clustering, Cluster Analysis, Density based Clustering, DBSCAN, Spatial Clustering, Spatial Data.