基于范数支持向量回归机的算法扰动设计研究

Perturbation Design of Algorithm Based on Norm Support Vector Regression Machine

解决回归问题中相对重要的方法就是支持向量回归机。实际问题中的一些固定测量及公式计算总会存在误差,因此有必要通过扰动分析来讨论研究支持向量回归机数据的扰动问题。基于此,把1范数支持向量回归机作为研究对象,通过扰动分析,在一定的条件下给出了解对扰动数据偏导数的表达式,建立线性规划中1范数支持向量机算法的原始问题的灵敏度分析定理。

Support vector regression machine is a relatively important method to solve regression problems.There are always errors in some fixed measurement and formula calculation in practical problems,so it is necessary for us to discuss and study the disturbance of SVM data through disturbance analysis.Based on this,taking 1 norm support vector regression machine as the research object,through disturbance analysis,the expression to understand partial derivatives of disturbance data is given under certain conditions,and the sensitivity analysis theorem of the original problem of 1 norm support vector machine algorithm in linear programming is established.