带有多面体扰动的半监督v-支持向量分类机

Semi-Supervisedv-Support Vector Machines with Perturbation in Polyhedron

在通常建立的优化模型中,一般都假定输入的数据是精确的,而实际生活中我们得到的数据总会带有测量或统计误差,因此,本文考虑数据在多面体内扰动的半监督两类问题,以v-支持向量分类机为基础,借鉴把半监督两类分类问题转化为一个凹规划的思想,给出数据在多面体内扰动的半监督v-支持向量分类算法。该算法的参数v易于选择,而数值试验也表明该算法具有良好的稳定性和较好的分类结果。

The classical paradigm in mathematical programming is to develop a model that assumes that the input data is precisely known and equal to some nominal values. In practice, the data usually have pertur- bations since they are subject to measurement or statistical errors. Therefore, we proposed the Semi-Supervisedv-Support Vector classification algorithm with perturbation in polyhedrons, which are based on formulating the problem as a concave minimization problem. It is solved by a successive linear approximation algorithm. Numerical experiments confirm that the parameter v is more stabile than parameter C, and the robustness of the proposed method.