DIAGNOSING STUDENT ERRORS IN E-LEARNING ENVIRONMENT USING MPE THEORY

摘要

An approach to trace and to analyze student actions in a web based learning environment (WBLE) is presented. An XML schema structuring relevant information in an e-session is defined inspired from SCORM recommendations. For each problem-solving activity for which a failure is noticed in a trace based on this schema, a two-level diagnosis procedure is used. Viewing the capability model associated with a test question or a problem solving activity as a Bayesian network, we show how belief revision theory provides a most probable explanation (MPE) for the observed failure.