基于数据挖掘的本体关系学习算法

Ontology Conceptual Relation Learning Algorithm Based on Data Mining

如何确定概念间语义关系的存在性和如何确定概念间的关系类型是本体关系学习的两个基本问题.现有的本体关系学习算法常常区分出不同类型的语义关系,使用不同的策略来获取概念间的各类关系,影响了算法的效率.提出一种基于数据挖掘的本体关系学习算法,运用关联规则挖掘获取概念间的关系,利用聚类分析对概念关系类型进行区分.实验结果证明,算法较好地解决了本体关系学习中的两个基本问题.

How to acquire the relation between the concepts and how to label the type of the conceptual relation are two basic problems of the ontology relation learning. Current algorithms often separate different concept relation types and use different methods to acquire ontology relations. Association rules mining and clustering analyses have been used in a large area such as retail, financial, communication and marketing organizations to find interesting data pattern. A data mining based ontology relation learning algorithm is proposed, which used association rules mining method to acquire relations between concepts from large text corpus and used clustering analyses method to distinguish the conceptual relations. According to the experimental result, the algorithm can resolve the two basic problems in ontology relation learning.