Noisy-free Length Discriminant Analysis with cosine hyperbolic framework for dimensionality reduction
作者：
Highlights：
• NRPS will remove noisy patterns with the help of a generalized threshold.
• NLDA models discrimination in terms of average length differences of the patterns.
• Cosine-hyperbolic framework formulates NLDA’s objective.
• This framework makes the length differences between samples more significant.
• NLDA is efficient in handling multimodal data.
摘要
•NRPS will remove noisy patterns with the help of a generalized threshold.•NLDA models discrimination in terms of average length differences of the patterns.•Cosine-hyperbolic framework formulates NLDA’s objective.•This framework makes the length differences between samples more significant.•NLDA is efficient in handling multimodal data.
论文关键词：Dimensionality reduction,Subspace learning,Discriminant analysis,Relevant patterns,Cosine hyperbolic,Face recognition
论文评审过程：Received 12 August 2016, Revised 14 March 2017, Accepted 15 March 2017, Available online 21 March 2017, Version of Record 30 March 2017.
论文官网地址：https://doi.org/10.1016/j.eswa.2017.03.034
