Document Type



A new algorithm ABC (Ant based Clustering) inspired from behavior of the real ants is proposed for clustering in data mining. ABC employs some ants clustering in the searching space. Different from some clustering algorithms, instead of picking or dropping data objects, artificial ants in ABC find and merge the similar data points in their vision range, also a following and randomly route choosing strategy is introduced. Computational simulation results show that ABC is efficient, and the clustering results can be achieved in short time and maintain stably.