Cooperative and penalized competitive learning with application to kernel-based clustering

作者：

Highlights：

• Propose a novel cooperative and penalized competitive learning (CPCL) model.

• Design the CPCL algorithm featuring high accuracy and performance robustness.

• Introduce the proposed competition mechanism in kernel-based clustering framework.

• Present a new kernel-based learning algorithm without knowing true cluster number.

摘要

Highlights•Propose a novel cooperative and penalized competitive learning (CPCL) model.•Design the CPCL algorithm featuring high accuracy and performance robustness.•Introduce the proposed competition mechanism in kernel-based clustering framework.•Present a new kernel-based learning algorithm without knowing true cluster number.

论文关键词：Competitive learning,Cooperation mechanism,Penalization mechanism,Kernel-based clustering,Number of clusters

论文评审过程：Received 19 September 2013, Revised 5 February 2014, Accepted 13 March 2014, Available online 21 March 2014.

论文官网地址：https://doi.org/10.1016/j.patcog.2014.03.010