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

Abstract

The present application relates to a spectral clustering method, device, system, computer equipment and storage medium. The method includes: sending the first encrypted target sample set to the second server, and the first encrypted target sample in the first encrypted target sample set is encrypted by the first public key; The first encrypted target similarity set obtained by the collection; the first encrypted target similarity set in the first encrypted target similarity set is decrypted by the first private key to obtain the first decrypted target similarity set; the encrypted first target similarity set is sent to the coordination server A target Laplacian matrix, the first target Laplacian matrix is ​​obtained according to the first decrypted target similarity set; receiving the encrypted clustering sample set sent by the coordination server according to the encrypted first target Laplacian matrix ; Perform joint clustering with the second server based on the encrypted clustering sample set. Adopting the method can improve the confidentiality of data.

Description

technical field [0001] The present application relates to the technical field of data processing, in particular to a spectral clustering method, device, system, computer equipment and storage medium. Background technique [0002] With the development of computer technology and the development of data processing technology, data clustering analysis is required in many application scenarios, and many clustering algorithms have emerged, such as Mean-Shift clustering algorithm and spectral clustering algorithm. Due to its universal applicability to data distribution and insensitivity to data sparsity, spectral clustering has become a mainstream clustering algorithm. [0003] At present, before clustering and analyzing the data, it is first necessary to obtain a large amount of valid data. However, the data of a single data owner is often limited. To obtain a large amount of data, it usually needs to come from multiple data owners. [0004] However, data is often confidential, a...

AI Technical Summary

Problems solved by technology

[0004] However, data is often confidential. When performing cluster analysis on confidential data, it is often necessary to provide confidential data to others, which causes data leakage and leads to low data confidentiality.

Images