论文部分内容阅读
针对传统文档自动分类方法和目前语义分类方法中存在的问题,提出一种新的基于概念向量空间的文档语义分类模型,该模型通过字符匹配算法将原文档高维词向量空间中相互独立的词项匹配到描述本体概念的属性集合,进而映射成属性集合对应的本体概念,形成低维的、语义丰富的文档概念向量空间。采用目前非常流行的数据集“20Newsgroups”作为实验数据集,对基于概念向量空间的文档语义分类模型进行实验验证。实验结果表明:提出的文档语义分类方法与传统基于词向量空间的文档分类方法相比,能够极大地降低向量空间维度,提高文档分类的性能。
Aiming at the problems existing in the traditional methods of automatic document classification and the current semantic classification methods, a new document semantic classification model based on concept vector space is proposed. The model matches the original documents The items are matched to the attribute set describing the concept of ontology and then mapped into the ontology concepts corresponding to the attribute set to form a low-dimensional and semantic rich document conceptual vector space. Using the currently very popular dataset “20Newsgroups ” as experimental dataset, the document semantic classification model based on concept vector space is validated experimentally. The experimental results show that the proposed document semantic classification method can greatly reduce the vector space dimension and improve the performance of document classification compared with the traditional document classification method based on word vector space.