engleski

DBSCAN-like clustering method for various data densities

In this paper, we propose a modification of the well-known DBSCAN algorithm, which recognizes clusters with various data densities in a given set of data points $A = {; ; a^i in R^n : i = 1, ldots , m}; ; $. First, we define the parameter $MinPts = floor ln |A| floor$ and after that, by using a standard procedure from DBSCAN algorithm, for each $a in A$ we determine radius $epsilon_a$ of the circle containing $MinPts$ elements from the set $A$. We group the set of all these radii into the most appropriate number $(t)$ of clusters by using Least Square distance-like function applying {; ; tt SymDIRECT}; ; or {; ; tt SepDIRECT}; ; algorithm. In that way we obtain parameters $epsilon_1 > · · · > epsilon_t$. Furthermore, for parameters ${; ; MinPts, epsilon_1}; ; we construct a partition starting with one cluster and then add new clusters for as long as the isolated groups of at least $MinPts$ data points in some circle with radius $epsilon_1$ exist. We follow a similar procedure for other parameters $epsilon_2, ldots, , epsilon_t$. After the implementation of the algorithm, a larger number of clusters appear than can be expected in the optimal partition. Along with defined criteria, some of them are merged by applying a merging process for which a detailed algorithm has been written. Compared to the standard DBSCAN algorithm, we show an obvious advantage for the case of data with various densities.

Clustering, DBSCAN, Incremental algorithm, Various data densities, Clusters merging, Least Squares distance-like function

nije evidentirano