Research on ant colony clustering algorithm based on HADOOP platform

Abstract

Due to in the early period of the ant colony clustering algorithm convergence speed is very slow, this paper proposes a hybrid clustering algorithm based on ant colony clustering and MMK-means algorithm, which uses MMK-means algorithm to process the data, followed by ant colony clustering to finish clustering. Apart from that, this paper improves the ant colony clustering algorithm that makes ants using the best matching position, data object placement selecting and so on. We realize the algorithm in Hadoop platform, which can effectively reduce the time costs of clustering.