An efficient Nyström spectral clustering algorithm using incomplete Cholesky decomposition

作者：

Highlights：

• A new matrix factorization strategy is designed for Nyström spectral clustering.

• Incomplete Cholesky decomposition is introduced to accelerate Nyström approximation.

• An efficient Nyström spectral clustering algorithm called ICD-NSC is proposed.

• The effectiveness of ICD-NSC is demonstrated by comprehensive experiments.

摘要

•A new matrix factorization strategy is designed for Nyström spectral clustering.•Incomplete Cholesky decomposition is introduced to accelerate Nyström approximation.•An efficient Nyström spectral clustering algorithm called ICD-NSC is proposed.•The effectiveness of ICD-NSC is demonstrated by comprehensive experiments.

论文关键词：Spectral clustering,Nyström approximation,Incomplete Cholesky decomposition,Large data set

论文评审过程：Received 8 April 2020, Revised 22 July 2021, Accepted 24 August 2021, Available online 31 August 2021, Version of Record 8 September 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115813