A model transformation from NL to SBVR

Abstract

In Requirement Engineering, requirements are usually written in sentences of natural language and natural languages are ambiguous and inconsistent, so the requirements written in natural languages also tend to be ambiguous. To avoid this problem of ambiguity we present an approach of model transformation to generate requirements based on SBVR (Semantics of Business Vocabulary and Business Rules). The information provided in source metamodel (NL) is automatically transformed into target metamodel (SBVR). SBVR metamodel can not only be processed by machine but also provides precise and reliable model for software design. The standard SBVR metamodel is already available but for natural language we proposed our own metamodel because there is no standard metamodel available for natural languages.