| With the development of industrial technology,electromechanical equipment has become diverse and complex.Workers in this field often need to retrieve a large amount of textual information such as equipment descriptions and maintenance manuals and only obtain a small part of them in their daily operation work.Internet resources are increasing.General search engines and professional forums have also become methods for electromechanical operation workers to retrieve and consult.However,these ways usually provide a large number of web pages containing irrelevant information,and users still need to work hard to screen.With the development of knowledge graph(KG)technology,more and more fields such as medical treatment and law begin to use knowledge graph to organize information,and use it as database to build efficient intelligent search,and question answering applications.The field of electromechanical operation can learn from this method to create an accurate,fast,and efficient intelligent search or question answering(QA)application.Aiming at the difficulty of data retrieval for workers in the electromechanical operation field,this paper provides an intelligent search solution based on the domain knowledge graph,and proposes a result ranking design that takes into account the individual needs characteristics of electromechanical operation workers,aiming to provide users with knowledge search results that are more in line with their current needs.The main work of this paper is as follows:(1)Construct a knowledge graph for electromechanical equipment operation:according to the analysis of the work content and knowledge data in this field,construct a schema with "equipment" as the central entity,and complete data collection and processing based on encyclopedia websites,electromechanical forums and others to obtain formatted knowledge,supplemented by manual label of profession,scope,and difficulty,and use graph database to construct and store of electromechanical equipment operation KG.(2)Design a QA model in the electromechanical equipment operation field that combines user characteristics: the design proposed in this paper not only considers the semantic analysis of the query question itself to obtain the target query,but also considers the user’s need characteristics from the three dimensions of major,scope and ability.When the question involves a wide range of knowledge,it can provide users with results that match their needs as much as possible.In the task of question semantic analysis,the rule-based method and the Bert-based deep learning method are combined to work,and realize the effective entity/relationship/attribute recognition on the problem set of the electromechanical field.(3)Develop an intelligent search system of the electromechanical equipment operation field: integrating the results of the above two tasks and using the Django framework for foreground-background development to complete an intelligent search system of electromechanical equipment operation,which can provide users with personalized intelligent search services within the scope of the knowledge base. |
