论文部分内容阅读
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.
This paper proposes the use of Newton’s method and conjugate gradient method to solve the nonlinear support vector regression learning problem, which can not only speed up the process of model selection but also improve the training speed.This method is applied to gas stove data set modeling and Mackey-Glass chaotic time series Prediction, simulation results show the effectiveness of the method.