AntPart: an algorithm for the unsupervised classification problem using ants

摘要

Unsupervised classification is one of the tasks of data-mining. In this paper, a method named AntPart for the resolution of exclusive unsupervised classification is introduced. It is inspired by the behavior of a particular species of ants called Pachycondyla apicalis. The performances of this method are compared with those of three other ones, also inspired by the social behavior of ants: AntClass, AntTree and AntClust.