Journal of Information Security Research Hamid Parvin, Hojjatollah Ahangarikiasari, Hamid Alinejad-Rokny 8 4 2017 http://www.dline.info/jisr/fulltext/v8n4/jisrv8n4_2.pdf Ensemble-based learning is a trustworthy option to reach a strong partition. In order to making up for the faults of each other, the classifiers in the ensemble can do the classification task more reliable than each of them. There is a straight way to induce a set of primary partitions that vary from each other, and then to gather the partitions through a gratifying functions to induce the final partition. Another option in the ensemble learning is to turn to fusion of different information from genuinely various sources. This article introduces a new clustering ensemble learning based on the Ant Colony clustering algorithm. Ensemble needs variability and swarm which is involved in randomness. Various runnings of ant colony clustering cause a number of diverse partitions. Considering these consequences as a new space datasets we make a final clustering by a simple partitioning algorithm to gather them in a gratifying partition. Experimental consequences on some real-world datasets are shown to present the effectiveness of the proposed method in inducing the final partition.