Automatic Extraction of SBVR Based Business Vocabulary from Natural Language Business Rules

In the early phases of software development, both of business analysts and IT architects collaborate to define the business needs in a consistent and unambiguous format before exploiting them to produce a software solution to the problem have been defined. Given the divergence of the interest areas of each intervenor, the natural language remains the most adequate format to define the business needs in order to avoid misunderstanding. This informal support suffers from ambiguity leading to inconsistencies, which will affect the reliability of the final solution. Accordingly, the Object Management Group (OMG) has proposed the “Semantic Business Vocabulary and Rules” (SBVR) standard which offers the opportunity to gather business rules in a natural language format having a formal logic aspect, letting the possibility to be understood by not only the different stakeholders but also directly processed by the machine. Since the SBVR standard is born to represent business rules by combining business vocabulary, it would be wise to give a great attention to the latter. In this paper we present an approach to extract business vocabulary according to SBVR Structured English as one of possibly notation that can map to the SBVR Meta-Model, with a view to provide a relevant resource for the next software deployment steps.