Spectral clustering acceleration method and system, computer equipment and storage medium

Abstract

The invention provides a spectral clustering acceleration method and system, computer equipment and a storage medium, solves the problem of long time consumption of excellent anchor point selection and K mean value calculation in a current spectral clustering method; the invention provides the spectral clustering acceleration method, the spectral clustering acceleration system, the computer equipment and the storage medium. According to a traditional approximate spectral clustering method, ZZT is constructed through a sparse representation matrix Z constructed by selected anchor points and original data points to approximately represent a Laplacian graph matrix, then a corresponding feature vector is obtained for carrying out K-means clustering so as to obtain a final clustering result matrix. However, according to the invention, final K-means clustering is not needed; when the scale of original data to be subjected to spectral clustering is relatively large actually, the time consumed by K-means clustering is long. According to the invention, the K mean value operation scale is changed from all points to anchor points, and the obtaining time of excellent anchor points is shortened; on the premise of ensuring a certain accuracy rate, the invention reduces the calculation time of a spectral clustering approximation algorithm, especially the problem of large-scale spectral clustering is solved, so that the operation time can be greatly reduced.

Description

technical field [0001] The present invention relates to the technical field of large-scale data clustering processing, and more specifically, to a spectral clustering acceleration method, system, computer equipment and storage medium. Background technique [0002] In recent years, with the development of artificial intelligence, machine learning, and computer vision, the research on data clustering methods has become particularly important. Data clustering has been widely used in the segmentation of medical images and the classification of financial data. Spectral clustering in data clustering is one of the most popular and important clustering methods in pattern recognition, machine learning and data mining. However, the high computational complexity of traditional spectral clustering methods hinders their application to large-scale datasets. Spectral clustering needs to construct a complete graph and perform eigendecomposition on this graph. For clustering problems with n...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problem that the selection of excellent anchor points and the calculation of K-means take a long time in the current spectral clustering method, the present invention proposes a spectral clustering acceleration method, system, computer equipment and storage medium to accelerate the clustering process and reduce the calculation time

Images