基于本体的故障案例信息抽取方法研究

以飞机维修保障中的经验知识积累和重用为目的,针对故障案例知识由于缺乏结构化、规范化描述,导致共享与重用困难的问题,对飞机故障案例的知识表达与信息抽取方法进行了研究.首先,根据飞机故障领域的特殊性以及知识共享和重用的实际需求,建立了飞机故障案例知识的本体模型;其次,利用中文分词工具以及文本工程通用框架(GATE),研究了对故障案例信息文档的语义标注以及基于规则的信息抽取技术;最后,利用Jena推理机挖掘出隐性信息,并实现在信息抽取过程中,通过不断发现新知识,主动扩展知识库.在此基础上开发了信息抽取原型系统,实现了从多种不同类型的文档信息中抽取出结构化故障案例信息,并利用数据库进行存储和管理,提高了故障案例知识的重用性,验证了研究方法的可行性.

Fault case information extraction method research based on ontology

To solve the accumulation and reusing problems of fault case knowledge that are described as unstructured and unnormalized information in the current maintenance support activities of aircraft, research on the knowledge representation and information extraction method of aircraft fault case was carried out. Firstly,ontology model of aircraft fault case knowledge was established according to the particularity of aircraft fault domain and the actual demand of knowledge sharing and reusing. Then with Chinese segmentation tools and general architecture for text engineering (GATE) frame, semantic annotation and rule based information extraction technology of aircraft fault case documents were studied. Finally, the hidden knowledge was discovered by using apache Jena inference engine, and knowledge base was expanded by the new knowledge found in the process of information extraction. Moreover, the prototype system for information extraction was developed and was used to extract structured fault case information from different types of documents, the information was then stored and managed by using database. This method was proved feasible to improve the reusability of fault case knowledge.