Cognitive modeling and the evolution of the student model in intelligent tutoring systems

Abstract

This thesis deals with the design and development of a student module for an intelligent tutoring system (ITS). Within the context of this thesis a student module has two components: a student model and a diagnostic component. We present an in-depth review of the history of ITS, on the design, development and limitations of the system. We review the methods of cognitive modeling used in some historical systems and a variety of methods of program analysis used in previous works. We also discuss some Ada language issues that are both language specific and computer specific, and related to tutoring in Ada. Our approach was to take an existing intelligent tutoring system, an Ada language tutor written in Ada, and extend the capabilities of the student module without altering the existing student model or diagnostic package. We designed heuristics to analyze the programming constructs involved in the student learning process and were able to generate information which may indicate the deficiencies and possible causes in the student model. The implementation of this idea required the integration of CLIPS and the Ada programming language. Using a small portion of the Ada language we have successfully diagnosed common programming misconceptions and missing concepts with novice Ada programmers.