Environment based clustering. A new approach
Laura Cristina Lanzarini, Armando Eduardo De Giusti · 2002
There are a variety of problems that require an automatic classification of a set of data. In this sense, clustering techniques have been widely applied, since they are known for forming classes or groups using a predefined similarity measure. This paper defines a new method which, as opposed to the solutions found so far, does not require any previous information about the data to be classified. The performance of this new proposal has been compared with a winner-take-all type method (WTA), which is widely used in clustering processes, and with the cluster-detection-and-labeling (CDL) method (Torbjorn, 1998), with satisfactory results.