【摘要】 支持向量回归是一种新的回归算法,通过非线性变换可完成原空间中的非线性回归。回归模型中参数的选择将直接影响到模型的概化能力。本文将精度参数的选择与训练样本及其噪音结合起来,给出了一种新的取值方法,数值试验验证了它的有效性和可行性。更多还原

【Abstract】 Support vector regression is a new regression algorithm and has good performance.It can effectively achieve nonlinear regression in input space via a nonlinear mapping.Parameter selection in the model will make great impact on its generalization performance.A new precision parameter selection method connected with distribution of training samples and their noise level was proposed in this paper,experiment testifies its validity and feasibility.更多还原