提出正则化框架下的半监督本体算法,将未标记数据的相关信息充分融入到计算模型中。同时,通过将数据嵌入到特征空间,利用特征表示从理论上得到特征空间上的正则化半监督本体学习算法。通过两个实验表明新算法对特定的应用领域具有较高的效率。
In this paper ,we pose semi-supervised ontology algorithm under the framework of the regularization such that the information of unlabeled data is fully integrated into the calculation model .At the same time ,the data is embedded in the feature space ,and feature space regularization semi-supervised ontology learning algorithm is given via the feature representation theory .Two experiments show that the new algorithm for specific applications with high efficiency .