摘要
本体映射通过计算实体之间的局部的或/和全局的语义相似度来确定实体之间的关系。除了实体之间的局部或/和全局的关系外,在很多情况下整个本体之间的关系对于计算语义相似度也是很有帮助的。本文试图利用这些关系来计算实体之间的相似度以改进本体映射的性能。为此,本文以范畴论为基础提出了一个层次化的本体模型(hierarchical ontology model,HOM)用以形式化描述这种关系。在此基础上,还提出了一种本体映射算法HOM-Matching,该算法利用本体之间的关系来计算实体之间的语义相似度。两组实验表明:通过设定合适的参数,该算法可以较好地提高本体映射的性能。
Ontology alignment establishes the relation between entities by computing their semantic similarities using the local or/and global context of entities. Besides the local and global context of entities, the relations between two ontologies are helpful for computing their semantic similarity in many situations. The aim of this article is to improve the performance of ontology alignment by using these relations in similarity computing. A hierarchical ontology model （HOM） which is based on category theory and describes these relations formally is proposed followed by HOM-matching, an algorithm based on HOM. It makes use of the relations between ontologies to compute semantic similarity. Two groups of experiments are conducted for algorithm validation and parameters optimization.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2009年第7期1750-1754,共5页
Systems Engineering and Electronics
关键词
本体
层次本体模型
本体映射
语义相似度
ontology
hierarchical ontology model
ontology mapping
semantic similarity
