A study on Quality Assessment of Requirement Engineering Document using Text Classification Technique

The Software Requirement Engineering document is the most important artifacts of the software development life cycle model. In majority of software systems, the Requirements Engineering (RE) Document or SRS (Software Requirement Specification) document has been written in natural language English that are prone to ambiguity. The ambiguous Requirement Engineering document may lead to disastrous results thereby hampering the entire development process and ending up compromising on the quality of a system. The success of any software product depends upon the quality of the Requirement Engineering document. The main reason for software crisis in software Industry is the Ambiguous Requirement Engineering Document. This paper discusses about the types of ambiguity, approaches to handle and providing a level of automatic assistance in order to detect ambiguity in the Requirement Engineering Document. The study also confirms the use of text classification technique to classify a text as “ambiguous” or “Unambiguous” at the syntax level. The key objectives of the work include understanding the presence of ambiguity in any Requirement Engineering document with the help of Machine Learning Techniques and finally minimizing or reducing it.