Survey of Kernel Extreme Learning Machine Kernel Function Based on the Perspective of Kernel Parameter Optimization Time

The Kernel Extreme Learning Machine (KELM) is more and more widely used in various fields, and the factor that determines the performance of KELM is the selected kernel function. The kernel functions are divided into two common types in this paper: the kernel inner product functions and kernel functions based on orthogonal polynomials that only require values in natural numbers. From the perspective of kernel parameter optimization time, the current research status and development of the above kernel functions are systematically summarized, which provided relevant researchers with the development history and latest results of the research direction.