A density-based algorithm for discovering clusters a density-based algorithm for discovering clusters in large spatial databases with noise

TLDR

Clustering algorithms are useful for class identification in spatial databases, but large datasets demand methods that require minimal domain knowledge, detect arbitrarily shaped clusters, and run efficiently—requirements unmet by existing algorithms. This paper introduces DBSCAN, a density‑based clustering algorithm designed to discover arbitrarily shaped clusters with minimal parameter tuning. DBSCAN uses a single input parameter and provides guidance for selecting it, and its effectiveness and efficiency were evaluated on synthetic data and the SEQUOIA 2000 benchmark. Experiments show DBSCAN outperforms CLARANS by more than 100‑fold in speed and is significantly more effective at finding arbitrarily shaped clusters.

Abstract

Clustering algorithms are attractive for the task of class identification in spatial databases. However, the application to large spatial databases rises the following requirements for clustering algorithms: minimal requirements of domain knowledge to determine the input parameters, discovery of clusters with arbitrary shape and good efficiency on large databases. The well-known clustering algorithms offer no solution to the combination of these requirements. In this paper, we present the new clustering algorithm DBSCAN relying on a density-based notion of clusters which is designed to discover clusters of arbitrary shape. DBSCAN requires only one input parameter and supports the user in determining an appropriate value for it. We performed an experimental evaluation of the effectiveness and efficiency of DBSCAN using synthetic data and real data of the SEQUOIA 2000 benchmark. The results of our experiments demonstrate that (1) DBSCAN is significantly more effective in discovering clusters of arbitrary shape than the well-known algorithm CLAR-ANS, and that (2) DBSCAN outperforms CLARANS by a factor of more than 100 in terms of efficiency.