Abstract:

A fuzzy C-means clustering algorithm based on semisupervised learning was proposed by introducing semisupervised learning into fuzzy C-means clustering algorithm. It has effectively solved the problem that the initial clustering centers random selection of fuzzy C-means algorithm can easily cause the local convergence and affects the clustering. The proposed algorithm can objectively obtain the optimal number of clusters and the initial cluster centers. Compared with the traditional FCM, our method can reduce the number of  iterations and the dependence on initial cluster centers.

Key words: semi-supervised learning, fuzzy C-means clustering algorithm, information entropy