Evolutionary clustering algorithm based on mixed measures

The purpose of this paper is to present an evolutionary clustering algorithm based on mixed measure for complex distributed data.

In this method, the data are first partitioned into some spherical distributed sub‐clusters by using the Euclidean distance as the similarity measurement, and each clustering center represents all the members of corresponding cluster. Then, the clustering centers obtained in the first phase are clustered by using a novel manifold distance as the similarity measurement. The two clustering processes in this method are both based on evolutionary algorithm.

Theoretical analysis and experimental results on seven artificial data sets and seven UCI data sets with different structures show that the novel algorithm has the ability to identify clusters efficiently with no matter simple or complex, convex or non‐convex distribution. When compared with the genetic algorithm‐based clustering and the K‐means algorithm, the proposed algorithm outperformed the compared algorithms on most of the test data sets.

The method presented in this paper represents a new approach to solving clustering problems of complex distributed data. The novel method applies the idea “coarse clustering, fine clustering”, which executes coarse clustering by Euclidean distance and fine clustering by manifold distance as similarity measurements, respectively. The proposed clustering algorithm is shown to be effective in solving data clustering problems with different distribution.