Fuzzy clustering based on distance metric under intuitionistic fuzzy environment

In the era of complex data, clustering techniques are useful in data mining. However, there are tremendous numbers of clustering techniques available in literature. Out of them, Fuzzy c-mean (FCM) technique is widely used due to its straightforwardness. However, FCM has some deficiencies such as initial sensitivity, easy fall into local minima and influencing through noise etc. To overcome such deficiencies, this study suggests an enhanced variant of the fuzzy c-means clustering technique within the framework of an intuitionistic fuzzy environment. The proposed algorithm reduces the effect of noise by developing a metric. To resolve the problem of uncertainty in assigning membership value, Sugeno's negation function is used to incorporate the hesitation degree. The proposed method is more flexible than previous methods due to the fact that it has the capability to deal with noise and uncertainty in assigning membership value. The feasibility and practicability of the proposed algorithm are demonstrated through its implementation on various benchmark data sets and conduct the comparative study with the extent methods. The experiment results state the admirable achievement of proposed algorithm over other tested algorithms.