An overview of incremental feature extraction methods based on linear subspaces
作者：
Highlights：
• Categorized overview of incremental feature extraction based on linear subspace methods.
• To add new information as it becomes available while retaining the previously acquired knowledge.
• Emphasis is done on methods with orthogonal matrix constraints based on global loss function.
• Incremental approaches are differentiated according to subspace model.
• Computational complexity, experimental setup and the performance are analyzed.
摘要
•Categorized overview of incremental feature extraction based on linear subspace methods.•To add new information as it becomes available while retaining the previously acquired knowledge.•Emphasis is done on methods with orthogonal matrix constraints based on global loss function.•Incremental approaches are differentiated according to subspace model.•Computational complexity, experimental setup and the performance are analyzed.
论文关键词：Incremental learning,Incremental feature extraction,Dimensionality reduction,Linear subspace methods,Covariance-based,Covariance-free
论文评审过程：Received 2 May 2017, Revised 19 January 2018, Accepted 20 January 2018, Available online 2 February 2018, Version of Record 20 February 2018.
论文官网地址：https://doi.org/10.1016/j.knosys.2018.01.020
