Study on improvement of alterable weighted FCM clustering algorithm based on genetic algorithm
Chen Xiao-guo · Journal of Science of Teachers' College and University · 2011
Fuzzy C—Mean(FCM)clustering is used widely,but it is easy to trap in local optimum and has sensitivity against the initial value.Definited the number and the chooseing method of the clustering centre with genetic algorithm at first,introduced weight coincidence in FCM,gave the recurrence formula and corresponding algorithm.The experimental results indicate that the algorithm can partly avoid the defect that FCM algorithm has sensitivity against the initial value and is easy to trap in local optimum.It make the clustering more reasonable and very well.