Comparative Analysis of Inductive Density Clustering Algorithms Meanshift and DBSCAN

The article presents an inductive model of objective clustering based on the MeanShift clustering technique. The algorithm for breaking an assortment of original data into two evenly powerful subsets is employed. The balance criterion is handled as an external criterion. To test the functioning of the proposed model, the “Jain” and “Flame” data sets from the Computing School of the East Finnish University were employed. The inductive DBSCAN algorithm was adopted for matching the preliminary outcomes. Based on the simulation proceeds, the ways for further improvement of the proposed model are arranged in order to increase the clustering objectivity of the examined data.