Multiple Kernel Extreme Learning Machine With Kernel Alignment Regularization

Abstract

Multiple kernel extreme learning machine (MKELM) is a research hotspot in the field of kernel learning. But we found that MKELM ignores the label information of samples when it optimizes the base kernel combination coefficient. In order to address this issue, the paper proposes multiple kernel extreme learning machine based on kernel alignment regularization, and the corresponding optimization method is given. The specific idea is that considering the sample label matrix as the ideal kernel in the kernel alignment criterion, and adding the criterion as a regularization term to the optimization goal of the multiple kernel extreme learning machine. Then, the corresponding optimization method is designed. The experimental results show that the proposed multiple kernel extreme learning machine based on kernel alignment regularization has good classification performance.