Multi-Objective Complete Fuzzy Clustering Approach

Abstract

AbstractThe process of data clustering has mainly focused on optimizing a single objective function, and thus, some information is not used for clustering. Therefore, the aim of this study is to propose a multi-objective complete fuzzy clustering model (MoCFC) that simultaneously optimizes data compactness, separation, and connectedness. The model employs two optimization algorithms; AUGMECON and NSGA-II. Using some fuzzy datasets, the results show that AUGMECON has lower convergence and coverage than NSGA-II, but a higher success index. Moreover, in terms of various cluster validity indices, AUGMECON achieves better performance. However, NSGA-II is the better choice if execution time is critical.