基于探测粒子群的小波核极限学习机算法

Wavelet Kernel Extreme Learning Machine Algorithm Based on Detecting Particle Swarm Optimization

在分析核极限学习机原理的基础上,将小波函数作为核函数运用于极限学习机中,形成小波核极限学习机(WKELM)。实验表明,该算法提高了分类性能,增加了鲁棒性。在此基础上利用探测粒子群(Detecting Particle Swarm Optimization,DPSO)对WKELM参数优化,最终得到分类效果较优的DPSO-WKELM分类器。通过采用UCI基因数据进行仿真,将该分类结果与径向基核极限学习机(KELM)、WKELM等算法结果进行比较,得出所提算法具有较高的分类精度。

In this paper,the principle of the kernel extreme machine was studied.Wavelet function was chosen to be the extreme learning machine's kernel function.Experiments show that this algorithm improves the classification accuracy and increases the robustness.Based on this method,we used detecting particle swarm optimization(DPSO)to optimize and set the initial parameters of WKELM in order to obtain the optimal WKELM classifier DPSO-WKELM.We used UCI gene data for simulation.The classification results are compared with the results of radial basis kernel extreme learning machine(KELM)and WKELM.The comparison shows that the proposed algorithm has higher classification accuracy.