Fuzzy System Inference and Fuzzy Cognitive Maps for a Cognitive Tutor of Algebra

Abstract

A cognitive tutor is an intelligent tutorial focused on supporting students on the resolution of problems. The large majority of the architectures for tutorial systems consider three common and indispensable modules: the student module; the tutor module; and, the domain module. These three modules are in charge of executing part of the more important activities within the tutor, and it is within these where it is necessary to implement models that simulate decision taking the way a human expert would. The student module should be constantly evaluating the progress made by students; the tutor module should provide adequate content, based on both the student's characteristics and the learning goals; and, the domain module should be capable of simulating the knowledge of experts regarding teaching-learning methodologies. Somehow, these three modules should interact in order to achieve the goals of the cognitive tutor. In this paper a proposal of fuzzy logic-based models for each of these modules is presented; the design of a system in which the three modules interact is also shown.