Spectral Clustering Algorithm Based on Density Representative Points

In order to solve the problems of long operation time and large memory consumption of the traditional spectral clustering algorithm applied to large-scale data sets, an improved spectral clustering algorithm based on density representative points is proposed. The algorithm changes the similarity matrix of spectral clustering. The construction method transforms the original spectral clustering from all sample points to construct a similarity matrix to use density representative points to construct a similarity matrix. In this way, the scale of the similarity matrix that needs to be calculated is greatly reduced, and the calculation of the spectral clustering algorithm is improved. In addition, the clustering effect of the spectral clustering algorithm is also improved. Simulation results show that the algorithm in this paper can effectively improve the processing ability of spectral clustering algorithm for data sets.