Bayesian system for student modeling

New technologies have provided the Education field with innovative aspects that allow significant improvements in teaching and learning processes. Their introduction not only reduces the effective cost of the application of pedagogical theories, but also opens up the possibility of exploring models from very different fields, facilitating their interaction and integration. One of the main innovations introduced since the first Computer Aided Learning programs are the so-called Intelligent Tutoring Systems (ITS), that, in contrast to traditional programs, have the ability to adapt to each individual learner. It is precisely this ability to adapt to each student what allows these programs to improve the teaching/learning process, as it has already been shown that the best learning method is individualized learning. Therefore, if the key characteristic of an ITS is its ability to adapt to each student, the key component of such a system is the student model, where all the information relative to the student is stored, including his/her cognitive state about the subject domain. The cognitive state is generated from student’s behavior during the interaction with the system, that is, it is inferred by the system from the information available: previous data about the student, answers to questions posed by the system, instructional episodes, etc. The process that consists in inferring the cognitive state of the student from observable data is called diagnosis. Diagnosis is without doubt the most complicated process in an ITS, since besides the inherent difficulty of any inference process it involves the treatment of information that in many cases is uncertain and/or imprecise. Frequently, ITS designers have preferred to develop their own heuristics instead of using Approximate Reasoning techniques available within the Artificial Intelligence field. The problem is that the lack of theoretical foundations of such heuristics makes the behavior of the system inadequate or unpredictable, yielding results different from the originally expected. The main goal of this dissertation is to improve the accuracy and efficiency of the diagnosis process in an ITS. To this end, we have explored the possibility of using Approximate Reasoning techniques, with special emphasis on simplifying their application as much as possible so their use is not an unaffordable load of additional work to the considerable task of developing an ITS. The proposed solution is substantiated in the definition of a new integrated student model based on Bayesian Networks (BN), and in the application of Computer Adaptive Tests (CAT) theory to improve the efficiency and accuracy of the diagnosis process. This student model allows substantial simplifications when defining the BN (nodes, links and parameters) used to represent the student model, and its integration with a CAT diagnostic algorithm (in which each question is selected adaptively according to the current estimation of student’s knowledge level) has shown an excellent performance in the evaluations we have realized with simulated students. The main original contributions of this thesis are: