BIRCHSCAN: UM MÉTODO DE APROXIMAÇÃO DO DBSCAN PARA GRANDES CONJUNTOS DE DADOS

Abstract

The DBSCAN algorithm is a classic density-based clustering method. This algorithm allows to identify clusters of different shapes, with the ability to identify noisy patterns in the data. DBSCAN presents good results, however it has a high computational complexity due to several distance calculations in the clustering process. This low computational efficiency limits its application to large data sets. This work presents a new method of grouping whose first stage is the construction of representative elements to apply DBSCAN to a reduced set of examples The proposed method allows grouping large datasets with approximate results to the DBSCAN result applied in the entire dataset. From the experiments performed, it is observed that the proposed technique presents good results and consistency when compared to other algorithms with a similar proposal.