A Novel K-harmonic Means Clustering Based on Enhanced Firefly Algorithm

摘要

The K-harmonic means is a center based clustering algorithm, which is suffering from falling into local optima easily. To solve this problem, a hybrid K-harmonic means algorithm using enhanced firefly algorithm is proposed. Combining with parallel chaotic optimization, a novel chaotic local search method which has the capability of full dimensional and part dimensional searching is used to enhance the original firefly algorithm. Then this enhanced version of firefly algorithm is integrated into K-harmonic means algorithm to take full advantage of merits of both algorithms. The proposed method is compared with KHM and other two hybrid algorithms on four data sets, and the experiment results show the superiority of it that can escape from local optima effectively.