Data Clustering Using Harmony Search Algorithm

Being one of the main challenges to clustering algorithms, the sensitivity of fuzzy c-means (FCM) and hard c-means (HCM) to tune the initial clusters centers has captured the attention of the clustering communities for quite a long time. In this study, the new evolutionary algorithm, Harmony Search (HS), is proposed as a new method aimed at addressing this problem. The proposed approach consists of two stages. In the first stage, the HS explores the search space of the given dataset to find out the near-optimal cluster centers. The cluster centers found by the HS are then evaluated using reformulated c-means objective function. In the second stage, the best cluster centers found are used as the initial cluster centers for the c-means algorithms. Our experiments show that an HS can minimize the difficulty of choosing an initialization for the c-means clustering algorithms. For purposes of evaluation, standard benchmark data are experimented with, including the Iris, BUPA liver disorders, Glass, Diabetes, etc. along with two generated data that have several local extrema.