Requirements to Class Model via SBVR

Abstract

A user's software requirements are represented in natural language or a speech such as English. Translating these requirements into the object-oriented models is a tough process for designers. This article proposes a neoteric approach to generate Unified Modeling Language (UML) class models instantly from software requirement specifications (SRS). Here the authors make use of the Open Natural language processing tool (OpenNLP) for lexical analysis and to generate the necessary parts of speech (POS) tags from these requirement specifications. Then, the Semantics of Business Vocabulary and Rules (SBVR) standard is used to extract the object-oriented elements from the natural language (NL) processed SRS. From this, the authors generate UML class models. The prototype tool can generate accurate models in less time. This automated system for designing object-oriented models from SRS reduces the cost and budget for both the designers and the users.