| With the improvement of equipments’ technical level,knowledge shortage has been a serious problem in maintenance and supporting work.Technical manual contains equipments’ structural,functional and operating knowledge,meanwhile diagnostic knowledge is implied in it.It is crucial to obtain these knowledge from technical manuals rapidly,and then to improve the efficiency or speed of maintenance action.As an improved semantic network approach to represent knoledge,conceptual graph has strong expressive ability and is easy to expand.vivid form and ability of extension,furthermore,reasoning process could be implemented through it.In this paper,a research based on conceptual graph to extract diagnostic knowledge in technical manuals is studied.The main content is as follows:(1)Introduce conceptual graph to extract diagnostic knowledge Use ontology theory to classify concepts and semantic relations,and then establish the domain ontology of technical manuals.Based on the domain ontology,a concept database including name entities and functional relations is created,which is used to extract concepts and label semantic labels.(2)Parse technical text using natural language processing tool,through which syntactic relations in sentences could be obtained.According to semantic relations and domain ontology,implement further analysis on syntactic relations and search the translating rules from syntactic relations to semantic relations.(3)Summarize translating rules and establish rule database.When the relation nodes between concepts need to be defined,match the rule in database and if all attributes could be matched to some rule,the semantic relation in the rule is the relation node.(4)Present a system to extact diagnostic knowledge based on conceptual graph and choose testing text to evaluate this system.The result indicates that the system has good performance in extract concepts and semantic relations,furthermore,reasoning process could be implemented on the built conceptual graph to diagnose fault. |
PDF Full Text Request