基于无约束优化的非线性支持向量回归

Nonlinear SVR based on unconstrained optimization

提出利用牛顿法以及共轭梯度法解决非线性支持向量回归学习问题,不仅可以加速模型选择的过程,而且能够提高训练速度.将该方法应用于煤气炉数据集建模以及Mackey-Glass混沌时间序列预测,仿真结果表明了该方法的有效性.

A learning strategy based on Newton method and conjugate gradient method is proposed in this paper to solve the nonlinear support vector regression （SVR） training problem, which is able to accelerate not only the model selection process but also the training speed. By applying it into gas furnace data set modeling and Mackey-Glass chaotic time series prediction, the simulation results indicate the effectiveness of the proposed learning strategy.