Chaotic Artificial Bee Colony for Text Clustering
Kusum Kumari Bharti, Pramod Kumar Singh · 2014
Text clustering is widely used for creating clusters of the digital documents. Selection of cluster centers plays an important role in creating clusters of the documents. In this paper, we use artificial bee colony algorithm (hereinafter referred to as ABC) to select an appropriate cluster centers for text documents. The ABC is a swarm intelligence based algorithm inspired by intelligent foraging behavior of real honey bees. The ABC provides good exploration of the search space at a cost of exploitation. To address this issue, we use the chaotic map as a local search paradigm to improve its exploitation capability. The proposed algorithm chaotic artificial bee colony (hereinafter referred to as ChABC) is tested on two benchmark text datasets namely Reuters-21,578 and Classic4, and the obtained results are compared with k-means clustering, ABC, and a recent variant of ABC namely gbest guided ABC (hereinafter referred to as GABC). The comparisons show that the ChABC offers the better clustering quality and faster convergence among all the competitive algorithms in all cases.