On unsupervised simultaneous kernel learning and data clustering
作者：
Highlights：
• A Joint framework for multiple kernel learning and nonnegative matrix factorization based clustering has been proposed.
• A novel eigenvalue maximization based multiple kernel learning approach has been presented.
• Sparsity metric based on l1-l2 norm has been utilized to find sparse matrix factors.
• A difference of convex formulation has been used for solving the underlying non-convex optimization problem.
• The convergence of the proposed algorithm to a stationary point has been established.
摘要
•A Joint framework for multiple kernel learning and nonnegative matrix factorization based clustering has been proposed.•A novel eigenvalue maximization based multiple kernel learning approach has been presented.•Sparsity metric based on l1-l2 norm has been utilized to find sparse matrix factors.•A difference of convex formulation has been used for solving the underlying non-convex optimization problem.•The convergence of the proposed algorithm to a stationary point has been established.
论文关键词：Clustering,Matrix factorization,Correlation analysis,Kernel learning
论文评审过程：Received 3 October 2019, Revised 26 May 2020, Accepted 23 June 2020, Available online 24 June 2020, Version of Record 12 July 2020.
论文官网地址：https://doi.org/10.1016/j.patcog.2020.107518
