Domain Requirements Engineering Knowledge Graph And Scenario Modeling Research

The requirement acquisition process is at the front of software development,which aims to get accurate,clear and consistent description of customer problems and corresponding solution limitations.At present,with the rapid development of big data,artificial intelligence and the Internet of things,the software field has begun to expand to multiple application fields,which leads to the direction and content of software requirement become more complex,and the accuracy and difficulty of demand acquisition also increase.However,in the process of using traditional requirement modeling methods,there is a knowledge gap between users and requirements modelers,which leads to the requirement modelers unable to quickly understand the business problems described by business personnel.At the same time,in the process of requirements modeling,lack of user participation,can't effectively model the real needs of users.In view of the above problems,firstly,this paper analyzes the process and related requirements analysis methods of domain engineering for reference to conduct in-depth research on the domain requirements scenario modeling method in model based on Systems Engineering(MBSE).At the same time,combined with the important measures of the current era of big data,based on the relationship between the various demand elements in the above methods,it is proposed to establish the domain demand engineering knowledge graph,comprehensively display the complex inclusion,association,derivation and other relationships between the various elements of the domain demand engineering,reduce the business knowledge gap between users and modelers.In the process of building domain requirement engineering knowledge graph.In this process,firstly,the requirement entities and the relationship between them in the MBSE based requirement analysis method are analyzed.In order to ensure the accuracy and integrity of the ontology,the advantages of seven steps and cyclic iteration method are integrated.The domain requirement engineering knowledge ontology is constructed manually.From the results of the construction,it can be found that the domain requirement knowledge ontology covers six core elements of requirement modeling and four key element relationships.Secondly,after the requirements knowledge acquisition,in the process of demand knowledge acquisition,based on the demand specification texts of 20 projects,Bi LSTM-CRF is adopted and ATT-Bi GRU is used to extract and analyze the requirement knowledge entity and the relationship between entities in the actual requirement specification text.Through relevant comparative experiments,the effectiveness of the above methods in the domain requirement engineering knowledge entity and relationship extraction is proved.Through the above model,the triple data of requirement knowledge in related fields are obtained and stored in Neo4 j database through corresponding processing,thus forming the domain requirement engineering knowledge base.Finally,the results show that 261 requirement knowledge entities and408 requirement entity relationships are obtained through requirement knowledge acquisition,which finally forms the demand knowledge graph of the project,thus proving the effectiveness of the above methods in building requirement engineering knowledge graph.