Parameters Selection of Kernel Based Extreme Learning Machine Using Particle Swarm Optimization

The generalization performance of kernel based extreme learning machine (KELM) with Gaussian kernel are sensitive to the parameters combination (C, γ). The best generalization performance of KELM with Gaussian kernel is usually achieved in a very narrow range of such combinations. In order to achieve optimal generalization performance, the parameters of KELM with Gaussian kernel were optimized by using particle swarm optimization (PSO) in this paper. To verify its effectiveness, the proposed method was tested on nine benchmark classification data sets compared with KELM optimized by Grid algorithm.