Global and local structure preserving sparse subspace learning: An iterative approach to unsupervised feature selection
作者：
Highlights：
• We propose a novel unsupervised subspace learning model for unsupervised learning.
• We propose a greedy algorithm to solve a combinatorial model.
• We propose an iterative algorithm to solve the relaxed continuous model.
• We establish a whole iterate sequence convergence result of the iterative algorithm.
• We conduct extensive experimental studies about the proposed algorithms.
摘要
Highlights•We propose a novel unsupervised subspace learning model for unsupervised learning.•We propose a greedy algorithm to solve a combinatorial model.•We propose an iterative algorithm to solve the relaxed continuous model.•We establish a whole iterate sequence convergence result of the iterative algorithm.•We conduct extensive experimental studies about the proposed algorithms.
论文关键词：Machine learning,Feature selection,Subspace learning,Unsupervised learning
论文评审过程：Received 15 June 2015, Revised 20 October 2015, Accepted 16 December 2015, Available online 24 December 2015, Version of Record 8 February 2016.
论文官网地址：https://doi.org/10.1016/j.patcog.2015.12.008
