Hybrid model for learner modelling and feedback prioritisation in exploratory learning

Abstract

Individual and/or hybrid AI techniques are often used in learning environments for well-structured domains to perform learner diagnosis, create and update a learner model and provide support at individual or group level. This paper presents a conceptual model that employs a synergistic approach based on Case-Based Reasoning (CBR) and Multicriteria Decision Making (MDM) components for learner modelling and feedback generation during exploration in an ill-defined domain of mathematical generalisation. The CBR component is used to diagnose what students are doing on the basis of simple and composite cases; simple cases represent parts of the models that the learners could possibly construct during an exploratory learning activity, while composite cases, which are assembled from simple cases, correspond to strategies that learners may adopt to construct their models. Similarity measures are used to identify how close/far are the learners from solutions pre-specified and stored in the knowledge base. This information is then fed into the MDM component that is responsible for prioritising types of feedback depending on the context. The operation of the two components and the effectiveness of the synergistic approach are validated through user scenarios in the context of an exploratory learning environment for mathematical generalisation.

Context-Dependent Feedback Prioritisation in Exploratory Learning Revisited

“…if there are n criteria and m alternatives, n C 2 pairwise comparisons are needed for the criteria and n * m C 2 pairwise comparisons are needed for the alternatives ( m C 2 comparisons per criterion). In previous work [12] we have used AHP alone to produce the prioritisation feedback. Due to the amount of pairwise comparisons, however, we have considered as criteria only the stage within the task and produce a pairwise comparisons for different combinations of learner characteristics.…”