摘要
为了对大规模军事知识资源分类管理,提出了一种基于知识地图的文本分类(KMTC)方法。将知识地图划分成层次社区结构,构建了知识地图的社区结构树,社区结构树的节点具有主题聚集属性,可作为文本按主题分类的依据;基于知识单元进行文本特征向量提取,并引入复杂网络的度中心度概念,计算文本对应的特征向量值。与传统文本分类方法相比,将收集整理的军事领域文本作为训练文本集合进行试验验证,提高了文本分类准确性。
To manage the large-scale military source,a knowledge map based text classification(KMTC)method is suggested.The knowledge map is divided into the hierarchical community structure to build the community tree,the node of the community structure tree has the topic aggregation attribute as the basis topic classification of the text.Based on the knowledge unit,the text eigenvector is extracted,the corresponding eigenvalue or the text is calculated by the degree centrality of the complex network.Compared with the traditional text classification method,the collected military text is tested as a set of the training text,improving the effectiveness of the text classification.
作者
李冰
陈奡
张永伟
LI Bing ;CHEN Ao ;ZHANG Yongwei(The 28th Research Institute of China Electronics Technology Group Corporation, Nanjing 210007, China)
出处
《指挥信息系统与技术》
2018年第1期92-95,共4页
Command Information System and Technology
关键词
文本分类
知识地图
社区结构
特征提取
text classification
knowledge map
community structure
feature extraction