An efficient automated design to generate UML diagram from Natural Language Specifications

Abstract

The foremost problem that arises in the Software Development Cycle is during the Requirements specification and analysis. Errors that are encountered during the first phase of the cycle migrate to other phases too which in turn results in the most costly process than the original specified process. The reason is that the specifications of software requirements are termed in the Nature Language Format. One can easily transform the requirements specified into computer model using UML. To minimize the errors that arise in the existing system, we have proposed a new technique that enhances the generation of UML models through Natural Language requirements, which can easily provide automatic assistance to the developers. The main aim of our paper is to focus on the production of Activity Diagram and Sequence Diagram through Natural Language Specifications. Standard POS tagger and parser analyze the input i.e., requirements in English language given by the users and extract phrases, activities, etc. from the text specifies. The technique is beneficial as it reduces the gap between informal natural language and the formal modeling language. The input is the requirements laid down by the users in English language. Some stages like pre-processing, part of speech (POs), tagging, parsing, phrase identification and designing of UML diagrams occur along with the input. The application and its framework is developed in Java and it is tested on by implementing on a few technical documents.