Font Size: 

K-harmonic means (KHM), has been introduced as one of the vital solutions to the classical K-means, which alleviates the sensitive dependence on the initial clusters conditions for the latter. However, it does retain the same deficiency as K-means: execution of KHM has a propensity to converge to local optima easily. In response to circumventing this problem, a new variant of KHM based on a recent nature inspired firefly search approach – specifically, the Firefly K-harmonic means (FAKHM) algorithm – is resorted to. Assessment analysis on several artificial and real life datasets demonstrates the superiority of the proposed FAKHM algorithm.