融合规则推理的工程知识图语义查询

Semantic Query for Engineering Knowledge Graphs Combining Rule Reasoning

摘要: 知识图由事实节点和有向边组成的三元组相互连接而成，本体为这些事实和有向边提供抽象层面的语义信息支持，能够对工程领域难以结构化的知识提供形式化的表达，高效的工程领域知识图语义查询是十分必要的。本文针对工程知识表示和语义信息查询实际需求，采用面向对象技术，设计和实现一个混合开放世界假设和封闭世界假设，融合OWL本体和SWRL规则，基于逆向推理的知识图语义查询方法。该方法采用SQWRL查询风格，扩展了其语义查询限制。最后通过加工特征自动识别实例，验证语义查询的可行性。

Abstract: The knowledge graph is composed of fact nodes and directed edges. Ontology provides abstract semantic information supports for these facts and directed edges. It can provide formal representation for the knowledge that is difficult to be structured in the engineering field. An efficient semantic query for knowledge graphs is very necessary. Aiming at the actual requirements of engineering knowledge representation and semantic information query, this paper designs and implements a hybrid open world consumption and closed world consumption semantic query method for knowledge graphs, which combines OWL ontology and SWRL rules, employs object-oriented technology, and bases on backward chained reasoning. This method adopts SQWRL query style and extends its semantic query limitation. Finally, the feasibility of semantic query is verified by an example of automatic recognition for machining features.