Discriminant analysis based on reliability of local neighborhood

作者：

Highlights：

• This paper integrated the intrinsic graph and penalty graph into a unified framework.

• DA-RoLN emphasizes the margin of different classes in low-dimensional space.

• The influence of noise can be reduced adaptively in dimension reduction process.

• DA-RoLN is easy to be solved and it have a low computational cost

摘要

•This paper integrated the intrinsic graph and penalty graph into a unified framework.•DA-RoLN emphasizes the margin of different classes in low-dimensional space.•The influence of noise can be reduced adaptively in dimension reduction process.•DA-RoLN is easy to be solved and it have a low computational cost

论文关键词：Dimensionality reduction,Discriminant analysis,Manifold learning,Graph learning,Adjacency factor

论文评审过程：Received 1 October 2020, Revised 18 February 2021, Accepted 20 February 2021, Available online 5 March 2021, Version of Record 20 March 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.114790