Data clustering using a modified fuzzy min-max neural network

In this paper, a modified fuzzy min-max (FMM) clustering neural network is developed. Specifically, a centroid computation procedure in embedded into the FMM clustering network to establish the cluster centroid of each hyperbox in the FMM structure. Based on the hyperbox centroids, the FMM clustering performance in undertaking data clustering problems is measured using the cophenetic correlation coefficient (CCC). A series of experimental studies using benchmark datasets is conducted. The CCC scores obtained are compared with those from other clustering algorithms reported in the literature. The empirical findings indicate the effectiveness of FMM with the centroid formation procedure for tackling data clustering tasks.