Learning representative exemplars using one-class Gaussian process regression
作者：
Highlights：
• A novel sparse Bayesian algorithm for learning exemplars is proposed.
• The proposed method automatically locates exemplars among similar observations.
• Applications to data representation and cluster analysis are provided.
• Theoretical generalization error bound for the method is provided.
摘要
•A novel sparse Bayesian algorithm for learning exemplars is proposed.•The proposed method automatically locates exemplars among similar observations.•Applications to data representation and cluster analysis are provided.•Theoretical generalization error bound for the method is provided.
论文关键词：Representative exemplars,One class Gaussian process regression,Support-based clustering,Automatic relevance determination,Kernel methods
论文评审过程：Received 23 August 2016, Revised 27 June 2017, Accepted 2 September 2017, Available online 11 September 2017, Version of Record 22 September 2017.
论文官网地址：https://doi.org/10.1016/j.patcog.2017.09.002
