Kernel meets recommender systems: A multi-kernel interpolation for matrix completion

作者：

Highlights：

• Propose a kernelized matrix completion framework via multi-kernel interpolation.

• Learn an effective low-dimensional representation in an infinite Hilbert space.

• Provide a feasible solution to make the raw input data linearly separable.

• Present an auto-weighted method for multi-kernel representation and fusion.

摘要

•Propose a kernelized matrix completion framework via multi-kernel interpolation.•Learn an effective low-dimensional representation in an infinite Hilbert space.•Provide a feasible solution to make the raw input data linearly separable.•Present an auto-weighted method for multi-kernel representation and fusion.

论文关键词：Recommender systems,Kernel learning,Multi-kernel learning,Matrix completion,Matrix interpolation

论文评审过程：Received 14 December 2019, Revised 9 August 2020, Accepted 1 December 2020, Available online 3 December 2020, Version of Record 5 December 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.114436