Developing a Self-regulation Environment in an Open Learning Model with Higher Fidelity Assessment

Abstract

Automation of pedagogical interventions in Model-tracing cognitive tutors (MTCT) strongly depends on chained paradigms like a proper modeling of the knowledge involved behind the student’s actions. Knowledge is tracked for inferring its degree of mastery that convey to a constructive learning process. In this paper is presented a methodology based on a probabilistic model for generating pedagogical interventions under a self-regulated environment. The foundations for developing it are explicitly detailed up to their implementation, passing through the modeling of the cognitive and meta-cognitive student knowledge. Probabilistic model is encoded in a Bayesian network topology that increases fidelity assessment by independently diagnosing degree of mastery of the relevant knowledge components and allowing a straightforward interpretation of the knowledge involved in a student’s actions. Moreover, it is also interwoven with other processes for inferring decisions that will influence in the way pedagogical interventions are generated and promoting a self-regulated behavior. Preliminary results to assess effectiveness of the proposed approach are also presented by implementing it in a MTCT called TITUS.