نوع المشرف

مشرف رئيسي

تاريخ الاشراف على الرسالة من

2014

الى

2017

اسم الطالب

Rawan Alheisa

ملخص الرسالة

The Software Development Life Cycle (SDLC) process is a continuous activity, which encompasses multiple phases. The most fundamental and essential phase in every SDLC is the requirement engineering phase. The final output from this phase represents a contract between the customer and the software engineer. It has been the most important and time-consuming phase since it can determine the success or the failure of the delivery of the software project. The requirements are written in natural language. Natural language has an ambiguous nature, and it is not fully standardized when it comes to requirements gathering and writing. The fact that the requirements are written in natural language leads to the conclusion that they might cause some confusion and misunderstanding. This will be shown and further explained later when the developer defines the table of Software Requirements Specification (SRS). For the reason, in this thesis, we have developed an Intelligent Software Requirement Analyzer (ISRA) methodology, based on an Artificial Intelligence (AI) technique, which uses the Artificial Neural Network (ANN) to deal with Natural Language Processing (NLP) applications. Our work’s core function is to tackle the natural language text intelligently and tokenize the requirements’ text. Ultimately, to have clear and understandable tokens. The proposed ISRA methodology results show that using it will significantly help, speed up, and enhance the generation of components of SRS. ISRA has been implemented using MATLAB® Integrated Development Environment (IDE), which offers flexible programming objects for developing Neural Networks (NN), as well as other essential objects and plug-in capabilities.