Clustering Approach Using Multiobjective Non-Dominated Sorting Teaching Learning Based Optimization with Kernel Fuzzy C-Means Algorithm (NSTLBO-KFCM)

Clustering has evolved over a period and has become more sensitive and precise with the outcomes. No clustering algorithm is suited for all types of data sets, but partitional clustering lays foundation for almost all-important clustering algorithms. Single solution can be achieved in case of single objective clustering, but in multiobjective clustering the solution becomes set of solution vector. Multiobjective clustering has made the pattern recognition possible in more than one dimension. The solution strategy of Multiobjective clustering is conceptualized on Pareto dominance. A multiobjective NSTLBO-KFCM is implemented in this paper using non-dominated sorting technique. The clustering algorithm is then compared with multiobjective non dominated sorting genetic algorithm third generation-based kernel fuzzy C-means (NSGAIII-KFCM) algorithm and multiobjective particle swarm optimization-based kernel fuzzy C-means (MPSO-KFCM) algorithm. The algorithm is also compared with non-dominated sorting teaching learning-based optimization with fuzzy C-means (NSTLBO-FCM) algorithm and the results show that NSTLBO-KFCM is superior clustering algorithm.