Nonnegative matrix factorization with combined kernels for small data representation

作者：

Highlights：

• Formulating a combined kernel with the newly defined fractional-power Gaussian kernel.

• A new NMF method for learning multi-granular non-linear representation of small data.

• A gradient decent algorithm for combined-kernel NMF with rigorous convergence proof.

• Face recognition experiments compared with ten matrix factorization methods.

摘要

•Formulating a combined kernel with the newly defined fractional-power Gaussian kernel.•A new NMF method for learning multi-granular non-linear representation of small data.•A gradient decent algorithm for combined-kernel NMF with rigorous convergence proof.•Face recognition experiments compared with ten matrix factorization methods.

论文关键词：Combined kernel,Nonnegative matrix factorization,Data representation,Face recognition

论文评审过程：Received 26 January 2022, Revised 19 June 2022, Accepted 11 July 2022, Available online 14 July 2022, Version of Record 21 July 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118155