Designing CBL systems for complex domains using problem transformation and fuzzy logic : a thesis presented in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Computer Science at Massey University, Palmerston North, New Zealand
Some disciplines are inherently complex and challenging to learn. This research attempts
to design an instructional strategy for CBL systems to simplify learning certain complex
domains. Firstly, problem transformation, a constructionist instructional technique, is used
to promote active learning by encouraging students to construct more complex artefacts
based on less complex ones. Scaffolding is used at the initial learning stages to alleviate
the difficulty associated with complex transformation processes. The proposed
instructional strategy brings various techniques together to enhance the learning
experience. A functional prototype is implemented with Object-Z as the exemplar subject.
Both objective and subjective evaluations using the prototype indicate that the proposed
CBL system has a statistically significant impact on learning a complex domain.
CBL systems include Learner models to provide adaptable support tailored to individual
learners. Bayesian theory is used in general to manage uncertainty in Learner models. In
this research, a fuzzy logic based locally intelligent Learner model is utilized. The fuzzy
model is simple to design and implement, and easy to understand and explain, as well as
efficient. Bayesian theory is used to complement the fuzzy model. Evaluation shows that
the accuracy of the proposed Learner model is statistically significant. Further, opening
Learner model reduces uncertainty, and the fuzzy rules are simple and resemble human
reasoning processes. Therefore, it is argued that opening a fuzzy Learner model is both
easy and effective.
Scaffolding requires formative assessments. In this research, a confidence based multiple
test marking scheme is proposed as traditional schemes are not suitable for measuring
partial knowledge. Subjective evaluation confirms that the proposed schema is effective.
Finally, a step-by-step methodology to transform simple UML class diagrams to Object-Z
schemas is designed in order to implement problem transformation. This methodology
could be extended to implement a semi-automated translation system for UML to Object