An ontology-based knowledge support system for requirements analysis : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Computer Science at Massey University, Manawatū, New Zealand
An Ontology-based Knowledge Support System for Requirements Analysis (OKSSRA) is proposed and
developed, in order to help requirements analysts obtain the preliminary business knowledge.
Requirements Engineering (RE) is a sub-discipline of Software Engineering (SE). It is
involved in the whole software life cycle from the very first step throughout the process of software
development. Thus, The performances of requirements analysts are crucial to RE outcomes since
requirements analysts bridge the communication gap between business stakeholders and development
However, normally, there is a knowledge gap between requirements analysts and business stakeholders,
especially when analysts work for out-sourcing contractors. The existence of this knowledge gap may
seriously lower analysts' efficiency of communicating with business stakeholders and hinder their
performances on preparing requirements documentation. Obviously, preliminary business knowledge
related to the project will help analysts to narrow down the knowledge gap and improved their
performances on preparing requirements documentation. Obviously, preliminary business knowledge related
to the project will help analysts to narrow down the knowledge gap and improve their performances.
However, based on our survey, there is no existing RE tools providing such knowledge support to
analysts. Therefore, we proposed and developed OKSSRA to help analysts obtain the preliminary business
knowledge for narrowing down the knowledge gap.
There are three key modules in OKSSRA: (i) a semantic similarity measure module, (ii) an ontology
mapping module, and (iii) an automatic use case generating module.
In the semantic similarity measure module of OKSSRA, we proposed and developed a new semantic
similarity measure utilising WordNet and Normalised Google Distance (NGD). In the new measure,
NGD will be used to calculate a unique length for each edge in the shortest path between two
candidate concepts in the WordNet graph. The semantic similarity measure enables our system (i)
to assign related concepts for a user's queries to extend the queries; and (ii) to identify the
related business processes from the business knowledge repository.
The ontology mapping module of OKSSRA employs a newly developed ontology mapping method based on MIMapper
(Kaza and Chen 2008). In this new ontology mapping method, our newly proposed semantic similarity
measure will be used to matching class names and to locate the most informative instances of their
class.The ontology mapping module enables our system (i) to update the ontology-based repositories
in the system, and (ii) to integrate the ontology-based repositories with other repositories.
In the OKSSRA module for automatically generating Use Cases, we propose a set of mapping rules for
the system to automatically generating Use Cases based on the information retrieved from business
processes. The set of mapping rules specified how the components of a business process are transformed
into use case elements, e.g., actors, goals, and steps of scenarios. With this module, our system
is able to generate essential Use Cases automatically using business processes retrieved from MIT
A set of three test use scenarios and a questionnaire has been carefully designed to evaluate the
efficiency and effectiveness of OKSSRA. The experimental results show that (i) our system is useful
for obtaining business knowledge, (ii) our system is more effective than existing system developed for
similar purpose, and (iii) our system is able to provide a pleasant user experiences.