一种改进的PSO-SVM算法

An Improved PSO-SVM Algorithm

粒子群优化(Particle swarm optimization, PSO)算法在整个训练数据集上搜寻支持向量机(Support vector machine, SVM)最优惩罚参数C和高斯核参数σ时会出现搜寻时间过长的问题。为了解决该问题,我们提出了一种基于距离配对排序(Distance pairing sorting, DPS)支持向量预选取的PSO-SVM算法(DPS- PSO-SVM)。该算法先将训练数据集进行DPS支持向量预选取构造一个支持向量候选集,然后利用PSO算法在支持向量候选集上对SVM参数寻优,最后将最优参数输入到SVM算法中对支持向量候选集进行训练。本文采用UCI数据库中的Breast Cancer数据和Banknote Authentication数据进行数值实验,结果表明该算法既能够缩短参数寻优时间,还能够保持PSO-SVM算法的高分类精度。

In this paper, we introduce an improved PSO-SVM algorithm based on distance pairing sorting support vector preselecting. When particle swarm optimization (PSO) searches the optimal penalty parameter C and kernel function parameter σ of SVM on the whole training data set, the search time will be too long. In order to solve this problem, this paper proposes that the training data set uses distance pairing sorting support vector preselecting to obtain a support vector candidate set, and then the PSO parameter optimization process is put on the support vector candidate set. This can save a lot of parameter optimization time. The Breast Cancer data and Banknote Authentication data in UCI database are used in numerical experiments. The results show that the method can not only reduce the time of PSO parameter optimization, but also get good classification accuracy.