BIRCH: an efficient data clustering method for very large databases
SIGMOD '96 Proceedings of the 1996 ACM SIGMOD international conference on Management of data
OPTICS: ordering points to identify the clustering structure
SIGMOD '99 Proceedings of the 1999 ACM SIGMOD international conference on Management of data
LOF: identifying density-based local outliers
SIGMOD '00 Proceedings of the 2000 ACM SIGMOD international conference on Management of data
Data Mining: Concepts and Techniques
Data Mining: Concepts and Techniques
A local-density based spatial clustering algorithm with noise
Information Systems
Hi-index | 0.01 |
A good approach in data mining is density based clustering in which the clusters are constructed based on the density of shape regions. The prominent algorithm proposed in density based clustering family is DBSCAN [1] that uses two global density parameters, namely minimum number of points for a dense region and epsilon indicating the neighborhood distance. Among others, one of the weaknesses of this algorithm is its un-suitability for multidensity data sets where different regions have various densities so the same epsilon does not work. In this paper, a new density based clustering algorithm, MSDBSCAN, is proposed. MSDBSCAN uses a new definition for core point and dense region. The MSDBSCAN can find clusters in multi-variant density data sets. Also this algorithm benefits scale independency. The results obtained on data sets show that the MSDBSCAN is very effective in multivariant environment.