针对现有的属性效应控制方法无法有效控制非线性回归建模的属性效应问题,基于间隔最大化和结构风险最小化原则,通过向SVR目标学习准则中施加等均值条件约束,提出等均值支持向量回归机(EM-SVR).所提出的方法具有较好的泛化能力,同时继承了EM-LS的良好性能.实验结果验证了所提出方法的有效性. Aiming at the problem that the existing attribute effect control method can not effectively control the attribute effect of nonlinear regression modeling, based on the principle of maximizing the interval and minimizing the structural risk, the equal mean support vector (EM-SVR) .The proposed method has better generalization ability and inherits the good performance of EM-LS.The experimental results verify the effectiveness of the proposed method.