To search, Click below search items.


All Published Papers Search Service


Pattern Clustering via Ants Colony, Ward Method, Kohonen Maps


Rosangela Villwock, Maria Teresinha Arns Steiner, Paulo Henrique Siqueira


Vol. 12  No. 6  pp. 81-92


The collective and self-organization behaviors of social insects have inspired researchers to reproduce them. Methods inspired by ants are a great promise for clustering problems. In the clustering algorithm based on Ants, patterns are spread out on a grid and each ant has assigned a pattern. The ants are responsible for picking, transporting and dropping patterns on the grid. After the algorithm converges, the recovery of clusters is done using the patterns’ positions on the grid. The purpose with this paper is to present changes and improvements to the Ant-based Clustering algorithm originally proposed by [2], hereinafter called proposed algorithm, evaluating its performance relatively to the Ward Method, the Kohonen Maps and the ACAM algorithm (Ant-based Clustering Algorithm Modified), proposed by [1]. The major changes in the proposed algorithm were: the introduction of a comparison between the probability of dropping a pattern at the position chosen randomly and the probability of dropping this pattern at its current position the introduction of an evaluation of the probability of a neighboring position when the decision to drop a pattern is positive and the cell in which the pattern should be dropped is occupied and the substitution of the pattern carried by an ant in case this pattern is not dropped within 100 consecutive iterations. To assess the performance of the proposed algorithm three real and public databases were used (?RIS, WINE and PIMA Indians Diabetes). The results showed superior performance of the proposed algorithm over the ACAM for two of the three databases and equality with other methods.


Data Mining, Metaheuristics, Ant-Based Clustering