A hybrid cuckoo search and K-means for clustering problem

摘要

Cuckoo search algorithm (CSA) is one of behavior algorithm which is effective to solve optimization problem including the clustering problem. Based on investigation, k-means is also effective to solve the clustering problem specially in fast convergence. This paper combines two algorithms, cuckoo search algorithm and k-means algorithm in clustering problem called FCSA. Cuckoo search is used to build the robust initialization, while K-means is used to accelerate by building the solutions. The result confirms that FCSA's computational time in ten datasets is faster than the compared algorithm.