K-medoids co-word clustering method based on CFSFDP
Aiming at the problems of unstable,incomplete,unfair and unclear clustering in co-word analysis,an improved method-K-medoids co-word clustering method based on CFSFDP is proposed.This method adopts the idea of automatically deter-mining the number of clusters and the clustering center in the CFSFDP algorithm,and uses the algorithm result as the initial cluster-ing center of the K-medoids algorithm to reduce the risk of falling into the local optimal solution.The experimental results show that compared with the traditional clustering method,our method is superior in performance and has better clustering effect.In addition,compared with other improved methods,our method is more suitable for co-word clustering and topic mining in unstructured text.
万方数据
