The Study And Application Of Spectral Clustering Based On Fuzzy Theory

In recent years,the application of spectral clustering become more and more extensive. Unlike the traditional clustering algorithms, the spectral clustering algorithm applies spectral graph theory to solve the clustering of non-convex sphere of sample spaces, so that they can be converged to global optimal solution. It constructs a simplified data space making use of the eigenvectors that not only reduces the dimension of data but also gives clearer distribution of data in the subspace.Firsty, in this paper, an overview of cluster analysis and several methods commonly used are outlined; the spectral clustering algorithm and its classification is introduced in detail, the challenges and problems this algorithm facing are also discussed. Then,some valuable directions for further research are proposed. Spectral graph theory is simply introduced and the spectral clustering algorithms are categorized according to rules of graph partition, and typical algorithms are studied emphatically, as well as their advantages and disadvantages are presented in detail; With the fuzzy theory, its application in the clustering is analyzed; Then several effective evaluation methods commonly used are discussed, and the application of various methods are analysed.Then a new semi-supervised clustering & adaptive algorithm is introduced in this paper to overcome the shortages of the most existing algorithms. The experiments show good results.Finally,Based on spectral clustering of ascertainable clustering number, combining fuzzy theory with spectral clustering algorithm, a fuzzy spectral clustering algorithm about multiple documents clustering is proposed in this paper. The method is about solving the problem such as a document belonging to multiple document clusters at the same time. The good results are shown in the simulation experiments.