International Journal of Advanced Research in Computer Science and Software Engineering (IJARCSSE)
Clustering is a major data mining technique for discovering trends in large databases. Outlier detection is a method that finds data objects that are inconsistent to the remaining data in the cluster. In this paper, the authors present DBSCAN algorithm which can deal with clusters of different densities and performs the clustering. It allows identifying clusters of arbitrary shapes and present outlier detection with stratification which ranks the different densities object. They show an implementation of this algorithm using Weka tool and present the data mining results.