支持向量机的多层动态自适应参数优化

Multi-layer adaptive parameters optimization approach for support vector machines

首先提出了基于多层动态自适应搜索技术的最小二乘支持向量机参数优化方法,然后采用最小二乘支持向量机对典型非线性控制系统的辨识进行了研究.辨识结果表明,最小二乘支持向量机可以用于非线性控制系统辨识,多层动态自适应搜索方法确定了最优支持向量机参数,从而获得精确的非线性控制系统辨识结果.

A novel least squares support vector machines (LS-SVM) for function estimationis is presented. And then, a hyper parameters and kernel parameters optimization approach called multi-layer adaptive best-fitting parameters search method is developed. According to different learning problem, the optimization approach can obtain appropriate LS-SVM parameters adaptively. Non-linear control system identification is studied using the improved LS-SVM. The results show that the optimization approach can induct best-optimized parameters for LS-SVM, and optimized LS-SVM provides excellent control system identification precision.