将近似支持向量回归机应用到多属性决策问题,提出基于近似支持向量回归机的多属性决策方法。该方法从决策问题本身出发,构造学习样本,再通过近似支持向量回归机拟舍出多属性效用函数,从而实现对方案的排序。与支持向量机相比该模型参数少,核函数无需满足Mercer条件,算法简单、可靠。最后通过算例表明方法的可行性与有效性。
A method for solving multiple attribute decision making (MADM) is proposed based on proximal support vector regression machine (PSVRM). The proposed method extracts learning samples from the MADM problem,estimates the multiple attribute utility function,and then sorts the alternatives. It has less number of parameter and is simple and reliable; the kernel does not need to satisfy the Mercer's condition. An example demonstrates its feasibility and availability.