基于特征向量的本体概念上下位关系验证  

Hyponymy verification of ontology concepts based on feature vectors

作　　者：田晓丹[1] 王庆林[1] 李原[1] 

机构地区：[1]北京理工大学自动化学院,北京100081

出　　处：《中南大学学报（自然科学版）》2013年第S2期351-354,共4页Journal of Central South University:Science and Technology

基　　金：创新方法工作专项项目(2011IM010300;2012IM010300)

摘　　要：提出一种基于概念对的语义特征和语境特征的特征向量来验证概念之间是否具有上下位关系的方法。从概念对的构词、共现、特征词出现数量、概念对在句中位置的距离出发,构建概念对上下位关系的特征向量,并使用支持向量机(SVM)进行训练和预测,从而实现本体关系的抽取。A new method based on feature vectors of semantic and context features to verify whether a pair of concepts has the hyponymy relation was represented.Based on word formation,the co-occurrence of concepts,the number of feature words and the location of concepts,feature vectors of hyponymy were built.Support vector machines(SVM) were used to train and predict data,and the relation extraction of ontology concepts was finally achieved.

关 键 词：本体学习 上下位关系 特征向量 SVM 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP391.1[自动化与计算机技术—控制科学与工程]