A new fuzzy C-means with priority
Jun Gao, Lili Xiang, Jiandong Wang · 2007
As traditional fuzzy C-means (FCM) has its shortages, we present an improved algorithm: M_FCM. The theory of fuzzy equivalency is used to deal with the original samples for getting the number of clusters and the original clustering center. By improving the clustering objective functions, the abilities of C-means are to handle isolated points and to show the significance of each dimension in the samples for clustering effect also enhanced. An experiment is finally conducted to make the MFCM clearer.