Research On Fuzzy C-mean Clustering Algorithm Based On ?-? Decision Graph

Fuzzy clustering is one of the important methods to divide data sets without labels.With the arrival of the era of big data,the amount of data increases exponentially.However,most of the data is not labeled.How to accurately classify these data and provide more accurate services for users has become the focus of research in today's society.Fuzzy c-means algorithm(FCM)is a fuzzy clustering algorithm based on objective function,which is compared with typical "hard clustering" algorithm(such as k-means algorithm).FCM algorithm calculates the membership degree of each sample to all classes,and obtains more reliable and accurate classification results.However,in the clustering process,FCM algorithm needs to manually determine the number of clusters and is sensitive to the initial clustering center.Problems such as clustering iteration,slow convergence speed and local optimal solution are easy to occur.To solve these problems,this thesis propose an algorithm that combines FCM algorithm with decision graph(DGFCM).Firstly,the decision graph is used to automatically select the clustering center and cluster number,and then FCM algorithm is used to realize clustering.Firstly,this thesis summarizes the research methods of improving FCM algorithm at home and abroad,and expounds the improvement of FCM algorithm in 5 directions according to the objective criterion function.Secondly,the case shows that the initial clustering center has a great influence on the iteration times and clustering results of FCM.Then the fuzzy c-means clustering algorithm based on-decision graph is introduced in detail.Finally,UCI real data set and artificial data set were used as experimental sample data,and traditional FCM clustering algorithm,fast density peak clustering algorithm(DPC)and DGFCM algorithm were used for clustering.Three clustering evaluation indexes were calculated and their clustering effects were compared.Comparison experiments show that the improved FCM algorithm has fast convergence speed and high precision.