Enforcing structure on textual use cases via annotation models

Textual use cases are commonly used to represent software requirements at initial stages. However in most of the cases, these documents are unstructured. In this paper, we present a linguistic engine for processing textual use cases and extract a structured model in terms of an annotation model out of these use cases. An annotation model of a use case can further be used to generate various UML requirements models, Business Process Models and ontology. The implementation details of Natural Language Processing (NLP) technique employed by us for the linguistic engine is described in this paper in detail. Also, we consider a corpus containing 123 use cases from real-life industrial projects within our company, and translate them into annotation models using our NLP technique. For evaluating the performance of conversion we use a few metrics and report some promising results for our linguistic engine.