K-harmonic means data clustering with imperialist competitive algorithm

Data clustering is one of the most important tasks of data mining. This paper aims to describe an integrated data clustering method based on Imperialist Competitive Algorithm (ICA) and K-Harmonic Means (KHM) algorithm. The proposed method is called ICA-KHM. KHM is a well-known clustering method and its main drawback is to converge to local optimums. Imperialist competitive algorithm is an evolutionary global search and optimization algorithm inspired by socio-political process of imperialistic competition. ICA has high convergence rate and can be used to solve optimization problems with multiple local minima. The proposed method combined the advantageous aspects of ICA and KHM in data clustering process. The proposed method is evaluated on five well-known datasets from different domains, and compared with KHM, and ICA. The experimental results indicate that the ICA-KHM provides better results than the two other methods.