Clustering Technique With Shared Nearest Neighbors

At present,spectral clustering has become one of the main clustering methods,but the similarity measure of traditional spectral clustering algorithms can not reveal the true clustering of not well-separated data sets,the spectral clustering algorithm based on closeness of shared nearest neighbors can effectively improve the problem and enhance the clustering quality.Therefore,this paper mainly studies the spectral clustering algorithm based on closeness of shared nearest neighbors,the specific work is as follows:(1)Due to the high computational time and space complexity of spectral clustering algorithm based on closeness of shared nearest neighbors,when dealing with large-scale and high-dimensional data,the time overhead is large and the cost is too expensive.The algorithm may be due to the system For reasons such as insufficient memory and invalidation,an incremental version of the algorithm is proposed to improve the performance of clustering.The basic idea is to first decompose the entire data set into its several subsets,and then the proposed spectral clustering algorithm guarantees its promising clustering performance by running on each subset in an incremental way.A lot of experiments on artificial data sets and simulation data sets here indicate the effectiveness of the proposed spectral clustering algorithm.At the same time,the algorithm has low time consumption,high clustering accuracy,and can effectively cluster the increasing data sets.(2)Because the traditional laplacian matrix is usually disturbed by noise or generated from biased samples from the underlying distribution,the calculated indicator vector is different from the truth vector.In this algorithm,the eigenvectors are calculated by laplacian matrix and partition level side information.In addition,gaussian kernel fuzzy c-means clustering are added to the spectral clustering algorithm based on closeness of shared nearest neighbors,it has the double advantages of kernel function and fuzzy mathematics,which can make up for the influence of hard partition of spectral clustering algorithm on clustering result.Experimental results verify the effectiveness of the proposed algorithm of gaussian kernel and partition constrained spectral clustering based on shared nearest neighbors.