Approach to Intuitionistic Fuzzy Clustering Based on Weighted Sample Sets

To improve performance of intuitionistic fuzzy clustering for large sample sets, the concepts of equivalent samples and weighted sample sets based on intuitionistic fuzzy sets is defined. Objective function of intuitionistic fuzzy C-means clustering algorithm is presented based on weighted sample sets. Iterative formulas of clustering centers and matrix of membership degrees are gotten by using Lagrange multiplier method. Initialization algorithm of clustering centers is given based on weighted sample sets to speed up the convergence rate. It is proved theoretically and experimentally that suitable value of parameter ξ used to defining equivalent samples not only generates almost equivalent clustering result with original set , but also improves performance of algorithm greatly.