Student Modeling for an Intelligent Tutoring System: Based on the Analysis of Answer Inconsistency

Students' answers are often correct, then incorrect next time. Such inconsistency influences the students' learning enormously. Thus, this article proposes a methodology that integrates fuzzy theory, Hasse diagrams, answer trees, and a layered structure to discover a student's answer inconsistency through scores for the tutor, to take effective strategies, in teaching the student. The main contributions of the article include (1) adopting fuzzy sets to model the imprecision of the real world, (2) extending the classification tree concepts to compute the score of partial correctness for identifying the students' misconceptions, and (3) employing Hasse diagrams to present the dynamic feature and reason the problem of incompleteness. The results of this article imply the systematical discovery of a student's answer inconsistency could obtain useful information for taking effective teaching strategies.