Models Derived from Automatically Analyzed Textual User Requirements

Requirements engineering is an important area of software engineering concerned with the extraction and presentation of knowledge from user requirements. There is a considerable gap between the various types of presentation of the same kinds of knowledge - those of the user written in natural language (NL) and those of requirements engineering (RE) depicted using diagrams (de facto standard UML diagrams, for example). Our research is aimed at filling that gap. In this paper, we introduce a universal formalism based on the basic building unit of NL, which is the relation triad. Through the definition of these basic relations in NL, we create three RE models: the use case path model, the hybrid activity diagram model and the domain model. These models are abstractions of the knowledge contained in the text, and serve as the basis for deriving UML diagrams