Research On Uncertainty Data Clustering Algorithm Based On Fuzzy Sets

With the rapid development of data mining technology,the application of clustering analysis technology has attracted more and more attention.Cluster analysis is a process of unsupervised learning.Clustering is clustered according to the similarity of data objects,and discovers the distribution and development trend of data in data sets.The unreliability,error,uncertainty,noise and other attributes in the real data set have a serious impact on the results of cluster analysis,therefore,research on the clustering of uncertain data has high practical value and can make the cluster analysis result more realistic.The uncertain data studied in this paper are mainly numerical and informative uncertain property data.In the clustering study of uncertain data,the clustering of uncertain data for processing obstacles and high-dimensional uncertain data are two unavoidable problems and two huge challenges.For the problems encountered in the above studies,the content of this paper is structured as follows:First of all,for the problem that the traditional uncertain clustering algorithm can not effectively solve the uncertain data of obstacles,this paper optimizes the existing obstacle-uncertainty clustering algorithm and proposes a Voronoi-based obstacle-space density clustering algorithm.This method introduces the triangular fuzzy number of Dev fuzzy concentration to solve the problem of data uncertainty.At the same time,it introduces the R-tree for pruning,reduces the computational complexity of the algorithm,and produces a relatively accurate data set.Then analyze the situation according to the obstacle constraints and improve the efficiency of the algorithm.Finally,the density clustering based on Voronoi diagram is used for cluster analysis to obtain more accurate and accurate clustering results.Secondly,in order to effectively cluster high-dimensional uncertain data,this paper uses projection subspace technology to reduce the dimension.By subspace projection of high-dimensional uncertain data,the impact of irrelevant or redundant attributes on high-dimensional clustering is effectively reduced,and reduce the amount of calculation.At the same time,the initial solution of the clustering algorithm is given by using the approximate skeleton theory to make up for the problem that the projection subspace is prone to fall into the local solution and avoid the local extremum of the clustering result;in addition,an uncertain fuzzy clustering algorithm is proposed based on intuitionistic fuzzy sets and relative entropy techniques.The relative degree of entropy is used to effectively measure the degree of difference among uncertain data samples,and the stability,comprehensiveness,and accuracy of the clustering results are guaranteed.Based on the above research on clustering of uncertain data,the paper concludes with a systematic summary,and looks forward to the future research direction of the paper,aiming to further study and research for the next academic research.