A New Clustering Approach based on K-means and Krill Herd Algorithm

Data clustering is a popular data analysis technique that divides a set of data into meaningful subsets (clusters) without any prior information. Krill Herd algorithm is a novel nature-inspired algorithm for solving optimization tasks. This article presents a new clustering algorithm based on krill herd and K-means algorithm. A local search strategy is used to avoid getting stock in local optima. The quality of proposed algorithm is evaluated on some UCI datasets. The experimental results show that the proposed method outperforms the other well-known algorithms such as k-means, PSO and ACO.