A vector quantization neural network model of partial supervision Fuzzy C-Means

This paper presents a novel version of partial supervision fuzzy c-means (FCM) algorithm. In order to avoid of achieving local minimum and to improve the computation efficiency, a clustering neural network is designed for the new partial supervision FCM algorithm. We also prove that clustering neural network designed is equivalent to the corresponding new partial supervision FCM algorithm. Meanwhile, the experiments are given out and the results show that the neural network of the new partial supervision FCM algorithm can easily find the global optimization solution, and thus is an effective approach.