The Department of Computer Science at the University of Cyprus cordially invites you to the PhD Defense entitled:

A Natural Language-based Methodology to Formalize and Automate the Requirements Engineering Process

Speaker: Marinos Georgiades
Affiliation: University of Cyprus, Cyprus
Category: PhD Defense
Location: Room 148, Faculty of Pure and Applied Sciences (FST-01), 1 University Avenue, 2109 Nicosia, Cyprus (directions)
Date: Tuesday, August 23, 2011
Time: 09:00-10:00 EET
Host: Constantinos Pattichis and Andreas Andreou (pattichi | aandreou AT

Effective implementation of an Information System requires a formal, understandable and time-saving requirements engineering (RE) methodology. Current approaches in RE fail to provide a specific, easily understood formalization of the major parts of the stages of requirements discovery, analysis and specification, and they usually result to a requirements document with ambiguities, redundancies and inconsistencies. The present dissertation introduces Natural Language Syntax and Semantics Requirements Engineering (NLSSRE), a compact and clear-cut methodology that is intended to formalize and automate a large part of the RE process, including discovery, analysis and specification of user requirements for the development of Information Systems, utilizing syntactic and semantic concepts of natural language (NL). NLSSRE is designed so that the analyst is guided in advance, through a step-by-step approach, what specific types of data, functions, business rules and functional conditions to use and search for, what questions to ask, in what specific way to analyze the answers to these questions and how to document them using formalized sentential requirement patterns. The formalized requirements are then easily transformed, with the use of specific rules, into diagrammatic notations, including class diagrams, data flow diagrams and use-case diagrams, as well as use-case and textual specifications, the latter following a certain Software Requirements Specifications (SRS) template. The formalization provided by NLSSRE is achieved with the aid of Natural Language elements, such as verbs, nouns, genitive case, adjectives and adverbials, while its automation is realized through the use of a dedicated CASE tool. To evaluate the effectiveness and efficiency of the proposed methodology, an experimental study was performed through which the NLSSRE methodology was compared to the classical Object-Oriented RE approach, by applying both of them in a real-life setting. The results were quite encouraging indicating that NLSSRE performed much better than the classical approach, as assessed by various objective quality metrics, such as completeness, correctness, modifiability and prioritization. The difference between the two approaches was also significant in regard to efficiency, where NLSSRE performed much faster than the classical one.

Short Bio:
Marinos Georgiades is a Ph.D. Candidate at the Department of Computer Science, University of Cyprus. He received his master’s in Information Management and bachelor’s in Computer Science, from the University of Sheffield in 2001 and the University of Cyprus in 2000, respectively. His research interests include Software Engineering and more specifically Requirements Engineering with emphasis on the use of Natural Language for the formalization and automation of software requirements elicitation, analysis and specification.

  Other Presentations Web:
  Colloquia Web: