Deterministic Multi-kernel based extreme learning machine for pattern classification

作者：

Highlights：

• Integration of deterministic and multiple kernel learning approach.

• Feature vectors are determined based on holistic and local appearance.

• Hidden layer parameters are analytically designed rather than random selection.

• Resultant Kernel function is linear combination of pre-specified kernels.

• Applicable for classification problems containing heterogeneous data.

摘要

•Integration of deterministic and multiple kernel learning approach.•Feature vectors are determined based on holistic and local appearance.•Hidden layer parameters are analytically designed rather than random selection.•Resultant Kernel function is linear combination of pre-specified kernels.•Applicable for classification problems containing heterogeneous data.

论文关键词：Pattern recognition,Feature extraction,GLCM,Multi-kernel,Deterministic learning

论文评审过程：Received 1 September 2019, Revised 7 November 2020, Accepted 30 May 2021, Available online 7 June 2021, Version of Record 19 June 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115308