Intelligent distance diagnosis of students' solutions. DIRCE: Diagnostic Interactive Relationship-Causality Engine.

Abstract

One of the important and most difficult tasks of open distance learning is the efficient diagnosis of students' solutions of given problems and exercises. Intelligent diagnostic systems and tools may essentially assist in this type of diagnosis by employing new methods and techniques from the fields of Artificial Intelligence and Cognitive Psychology. In this paper we present DIRCE (Diagnostic Interactive Relationship-Causality Engine), a novel intelligent diagnostic engine suitable for computer-based distance diagnosis. The scope of DJRCE is the detection of discrepancies in the students' solutions, and the identification of relevant errors. DIRCE diagnoses problems solved in a procedural manner with steps interconnected through mathematical equations. The engine does not need extended student models or bug libraries. It is mainly based on the correct solution(s) of the problem, the interconnections among the steps of the student's solution, as well as on limited evidential information about the solution under examination acquired interactively from the user. DIRCE has been used in the DIAS and I-DIAS experimental systems diagnosing the students' solutions of problems in high school physics and chemistry.