Dealing with noise in ant-based clustering

Separating the noise from data in a clustering process is an important issue in practical applications. Various algorithms, most of them based on density functions approaches, have been developed lately. The aim of this work is to analyze the ability of an ant-based clustering algorithm (AntClust) to deal with noise. The basic idea of the approach is to extend the information carried by an ant with information concerning the density of data in its neighborhood. Experiments on some synthetic test data suggest that this approach could ensure the separation of noise from data without significantly increasing the algorithm´s complexity.