A new intuitionistic fuzzy c-means clustering algorithm

This paper presents a new intuitionistic fuzzy c-means (IFCM)clustering algorithm by adapting a new method to calculate the hesitation degree of data point in cluster. From the definition of fuzzy entropy, if a clustering result of a data point has bigger fuzzy entropy, the clustering result should have more uncertainty. It means that we have insufficient information to deal with the clustering of a data point, so the hesitation degree of clustering result of the data point should be greater. Form this opinion, a mathematical model is applied to calculate the hesitation degree of clustering of data point based on fuzzy entropy is given. An IFCM clustering algorithms is present. Experiments are performed using two-dimensional synthetic data-sets referred from previous papers. Results have shown that proposed algorithm is not only effective for linear and nonlinear separation, but also able to describe more information comparing to fuzzy c-means clustering algorithm.