中文摘要：

本文通过引入拉普拉斯（Laplacian）正则项,针对半监督分类问题我们建立了基于拉普拉斯正则项的半监督不平行超平面分类机。和经典的双支持向量机相比,该算法不仅继承了不平行超平面决策的优点,并且将其推广到了半监督分类问题中。最后在人工数据上进行数值实验,与拉普拉斯双支持向量机和拉普拉斯支持向量机做比较,数值结果表明我们提出算法的可行性和有效性,特别是对于交叉型数据集,基于拉普拉斯正则项的半监督不平行超平面分类机具有明显较高的分类精确度。

英文摘要：

In this paper, we have proposed a novd Laphcian nonparallel hyperphnes chssifier for the semi-supervised classification problem.Compared with the twin support vector machine, it has the advantage ofnonparalld hyperphnes classifier and can be used for the semi-supervised classification problean. Finally, compared with the Laphcian twin support vector machine and the Laphcian support vector machine, the results of experiments on artificial dataset and UCI datasets show that our method is feasible, especially for＂Cross Planes＂datasets.