基于蚁群聚类算法的SVM半监督式训练方法

传统支持向量机在处理包含大量未知类别样本的训练集时性能较差。针对这一不足,在少量已知类别样本和大量未知类别样本构成的训练集上,提出一种基于蚁群聚类算法的支持向量机半监督式学习方法。该方法应用蚁群聚类算法进行聚类分析,实现了同类样本的自组织聚类;通过一个递归的类别判定算法,回收样本类别;同时,提取各类簇之间靠得相对较近的边界样本组成精简训练集,以缩减训练集规模加快学习速度。实验表明,该算法能够自适应样本类别分布,有较高的分类精度和泛化能力。

Semi-supervised Training Approach of SVM Based on Ant Clustering Algorithm

To overcome the shortcomings of traditional SVM such as incompetence to manage training data set with large-scale unlabeled examples,a Semi-supervised training approach of SVM based on ant clustering algorithm is put forward.First of all,the ant clustering algorithm is applied to a large-scale training set consisting of a small number of labeled samples and a large number of unlabeled samples,and self-organization cluster within the similar samples is implemented.Then,the samples' class information can be p...