Research on Fuzzy Clustering Based on Subtractive Clustering and Fuzzy c-means
Min Zhang · Jisuanji gongcheng · 2005
It is very sensitive to its initial value when one uses fuzzy c-means (FCM) to put up fuzzy clustering. It will get into local optimum solution if the enactment of initial value is not good. So the paper uses subtractive clustering to initialize the initial value of FCM before it uses FCM to put up fuzzy clustering. If one does these,it will gain the optimum solution,quicken the rate of convergence and need not to give the cluster number beforehand.