改进的PSVM及其在非平衡数据分类中的应用

Modified PSVM and Its Application in Unbalanced Data Classification

标准近似支持向量机(PSVM)没有考虑非平衡分布数据的分类问题,为此,在PSVM的基础上,将优化问题中的惩罚因子由数值变更为一个对角阵,提出了一种改进的PSVM算法。该方法利用引入的对角阵对正负样本分别分配不同的惩罚因子,由于其任意性,使得该算法可以解决由多种因素引起的分布不平衡的分类问题,稳健性较好。利用实值免疫克隆算法实现了模型参数的自动选择,进一步提高了算法的泛化性能。实验结果表明新算法对于处理分布不平衡数据的分类问题相当有效。

A modified proximal support vector machine is presented for the case of unbalanced data classification in many applications. The algorithm assigns the different penalty coefficients to the positive and negative samples respectively by adding a new diagonal matrix in the primal optimization problem. And further the decision function is obtained. In addition, the real-coded immune clone algorithm is employed to select the global optimal parameters to get the high generalization performance. The experimental results on the UCI and the real radar datasets （HRRP） illustrate the effectiveness of the proposed method.